3062 lines
2.3 MiB
Plaintext
3062 lines
2.3 MiB
Plaintext
|
{
|
|||
|
|
"cells": [
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"# Приступаем к работе...\n",
|
|||
|
|
"\n",
|
|||
|
|
"### Начнём анализ данных. Выгрузим в датафрейм первый набор (Цены на бриллианты)\n",
|
|||
|
|
"\n",
|
|||
|
|
"P.S. Кааажется будет трудно... :|"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 1,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"name": "stdout",
|
|||
|
|
"output_type": "stream",
|
|||
|
|
"text": [
|
|||
|
|
"Index(['Unnamed: 0', 'carat', 'cut', 'color', 'clarity', 'depth', 'table',\n",
|
|||
|
|
" 'price', 'x', 'y', 'z'],\n",
|
|||
|
|
" dtype='object')\n",
|
|||
|
|
"\n",
|
|||
|
|
"<class 'pandas.core.frame.DataFrame'>\n",
|
|||
|
|
"RangeIndex: 53943 entries, 0 to 53942\n",
|
|||
|
|
"Data columns (total 11 columns):\n",
|
|||
|
|
" # Column Non-Null Count Dtype \n",
|
|||
|
|
"--- ------ -------------- ----- \n",
|
|||
|
|
" 0 Unnamed: 0 53943 non-null int64 \n",
|
|||
|
|
" 1 carat 53943 non-null float64\n",
|
|||
|
|
" 2 cut 53943 non-null object \n",
|
|||
|
|
" 3 color 53943 non-null object \n",
|
|||
|
|
" 4 clarity 53943 non-null object \n",
|
|||
|
|
" 5 depth 53943 non-null float64\n",
|
|||
|
|
" 6 table 53943 non-null float64\n",
|
|||
|
|
" 7 price 53943 non-null int64 \n",
|
|||
|
|
" 8 x 53943 non-null float64\n",
|
|||
|
|
" 9 y 53943 non-null float64\n",
|
|||
|
|
" 10 z 53943 non-null float64\n",
|
|||
|
|
"dtypes: float64(6), int64(2), object(3)\n",
|
|||
|
|
"memory usage: 4.5+ MB\n"
|
|||
|
|
]
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"import pandas as pd\n",
|
|||
|
|
"import matplotlib.pyplot as plt\n",
|
|||
|
|
"import seaborn as sns\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Первый датафрейм о бриллиантах\n",
|
|||
|
|
"df = pd.read_csv(\".//static//csv//diamonds_prices2022.csv\")\n",
|
|||
|
|
"print(df.columns)\n",
|
|||
|
|
"\n",
|
|||
|
|
"print()\n",
|
|||
|
|
"\n",
|
|||
|
|
"df.info()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"**Описание набора:** В этом наборе данных представлен датасет о ценах на бриллианты. Рассматривается набор данных, содержащий цены и атрибуты примерно для 54 000 бриллиантов круглой огранки. В наборе данных 53 940 бриллиантов с 10 характеристиками (карат, огранка, цвет, прозрачность, глубина, таблица, цена, x, y и z). Большинство переменных являются числовыми по своей природе, но переменные cut, color и clearity являются упорядоченными факторными переменными со следующими уровнями. О валюте для столбца price: это цена ($). Что касается столбцов x, y и z, то они представляют собой размеры алмаза в виде (( x: длина в мм, y: ширина в мм, z: глубина в мм))"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 17,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"text/html": [
|
|||
|
|
"<div>\n",
|
|||
|
|
"<style scoped>\n",
|
|||
|
|
" .dataframe tbody tr th:only-of-type {\n",
|
|||
|
|
" vertical-align: middle;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"\n",
|
|||
|
|
" .dataframe tbody tr th {\n",
|
|||
|
|
" vertical-align: top;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"\n",
|
|||
|
|
" .dataframe thead th {\n",
|
|||
|
|
" text-align: right;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"</style>\n",
|
|||
|
|
"<table border=\"1\" class=\"dataframe\">\n",
|
|||
|
|
" <thead>\n",
|
|||
|
|
" <tr style=\"text-align: right;\">\n",
|
|||
|
|
" <th></th>\n",
|
|||
|
|
" <th>Unnamed: 0</th>\n",
|
|||
|
|
" <th>carat</th>\n",
|
|||
|
|
" <th>cut</th>\n",
|
|||
|
|
" <th>color</th>\n",
|
|||
|
|
" <th>clarity</th>\n",
|
|||
|
|
" <th>depth</th>\n",
|
|||
|
|
" <th>table</th>\n",
|
|||
|
|
" <th>price</th>\n",
|
|||
|
|
" <th>x</th>\n",
|
|||
|
|
" <th>y</th>\n",
|
|||
|
|
" <th>z</th>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" </thead>\n",
|
|||
|
|
" <tbody>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>0</th>\n",
|
|||
|
|
" <td>1</td>\n",
|
|||
|
|
" <td>0.23</td>\n",
|
|||
|
|
" <td>Ideal</td>\n",
|
|||
|
|
" <td>E</td>\n",
|
|||
|
|
" <td>SI2</td>\n",
|
|||
|
|
" <td>61.5</td>\n",
|
|||
|
|
" <td>55.0</td>\n",
|
|||
|
|
" <td>326</td>\n",
|
|||
|
|
" <td>3.95</td>\n",
|
|||
|
|
" <td>3.98</td>\n",
|
|||
|
|
" <td>2.43</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>1</th>\n",
|
|||
|
|
" <td>2</td>\n",
|
|||
|
|
" <td>0.21</td>\n",
|
|||
|
|
" <td>Premium</td>\n",
|
|||
|
|
" <td>E</td>\n",
|
|||
|
|
" <td>SI1</td>\n",
|
|||
|
|
" <td>59.8</td>\n",
|
|||
|
|
" <td>61.0</td>\n",
|
|||
|
|
" <td>326</td>\n",
|
|||
|
|
" <td>3.89</td>\n",
|
|||
|
|
" <td>3.84</td>\n",
|
|||
|
|
" <td>2.31</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>2</th>\n",
|
|||
|
|
" <td>3</td>\n",
|
|||
|
|
" <td>0.23</td>\n",
|
|||
|
|
" <td>Good</td>\n",
|
|||
|
|
" <td>E</td>\n",
|
|||
|
|
" <td>VS1</td>\n",
|
|||
|
|
" <td>56.9</td>\n",
|
|||
|
|
" <td>65.0</td>\n",
|
|||
|
|
" <td>327</td>\n",
|
|||
|
|
" <td>4.05</td>\n",
|
|||
|
|
" <td>4.07</td>\n",
|
|||
|
|
" <td>2.31</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>3</th>\n",
|
|||
|
|
" <td>4</td>\n",
|
|||
|
|
" <td>0.29</td>\n",
|
|||
|
|
" <td>Premium</td>\n",
|
|||
|
|
" <td>I</td>\n",
|
|||
|
|
" <td>VS2</td>\n",
|
|||
|
|
" <td>62.4</td>\n",
|
|||
|
|
" <td>58.0</td>\n",
|
|||
|
|
" <td>334</td>\n",
|
|||
|
|
" <td>4.20</td>\n",
|
|||
|
|
" <td>4.23</td>\n",
|
|||
|
|
" <td>2.63</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>4</th>\n",
|
|||
|
|
" <td>5</td>\n",
|
|||
|
|
" <td>0.31</td>\n",
|
|||
|
|
" <td>Good</td>\n",
|
|||
|
|
" <td>J</td>\n",
|
|||
|
|
" <td>SI2</td>\n",
|
|||
|
|
" <td>63.3</td>\n",
|
|||
|
|
" <td>58.0</td>\n",
|
|||
|
|
" <td>335</td>\n",
|
|||
|
|
" <td>4.34</td>\n",
|
|||
|
|
" <td>4.35</td>\n",
|
|||
|
|
" <td>2.75</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" </tbody>\n",
|
|||
|
|
"</table>\n",
|
|||
|
|
"</div>"
|
|||
|
|
],
|
|||
|
|
"text/plain": [
|
|||
|
|
" Unnamed: 0 carat cut color clarity depth table price x y \\\n",
|
|||
|
|
"0 1 0.23 Ideal E SI2 61.5 55.0 326 3.95 3.98 \n",
|
|||
|
|
"1 2 0.21 Premium E SI1 59.8 61.0 326 3.89 3.84 \n",
|
|||
|
|
"2 3 0.23 Good E VS1 56.9 65.0 327 4.05 4.07 \n",
|
|||
|
|
"3 4 0.29 Premium I VS2 62.4 58.0 334 4.20 4.23 \n",
|
|||
|
|
"4 5 0.31 Good J SI2 63.3 58.0 335 4.34 4.35 \n",
|
|||
|
|
"\n",
|
|||
|
|
" z \n",
|
|||
|
|
"0 2.43 \n",
|
|||
|
|
"1 2.31 \n",
|
|||
|
|
"2 2.31 \n",
|
|||
|
|
"3 2.63 \n",
|
|||
|
|
"4 2.75 "
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"execution_count": 17,
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "execute_result"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Для наглядности\n",
|
|||
|
|
"df.head()\n"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 18,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"text/html": [
|
|||
|
|
"<div>\n",
|
|||
|
|
"<style scoped>\n",
|
|||
|
|
" .dataframe tbody tr th:only-of-type {\n",
|
|||
|
|
" vertical-align: middle;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"\n",
|
|||
|
|
" .dataframe tbody tr th {\n",
|
|||
|
|
" vertical-align: top;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"\n",
|
|||
|
|
" .dataframe thead th {\n",
|
|||
|
|
" text-align: right;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"</style>\n",
|
|||
|
|
"<table border=\"1\" class=\"dataframe\">\n",
|
|||
|
|
" <thead>\n",
|
|||
|
|
" <tr style=\"text-align: right;\">\n",
|
|||
|
|
" <th></th>\n",
|
|||
|
|
" <th>Unnamed: 0</th>\n",
|
|||
|
|
" <th>carat</th>\n",
|
|||
|
|
" <th>depth</th>\n",
|
|||
|
|
" <th>table</th>\n",
|
|||
|
|
" <th>price</th>\n",
|
|||
|
|
" <th>x</th>\n",
|
|||
|
|
" <th>y</th>\n",
|
|||
|
|
" <th>z</th>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" </thead>\n",
|
|||
|
|
" <tbody>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>count</th>\n",
|
|||
|
|
" <td>53943.000000</td>\n",
|
|||
|
|
" <td>53943.000000</td>\n",
|
|||
|
|
" <td>53943.000000</td>\n",
|
|||
|
|
" <td>53943.000000</td>\n",
|
|||
|
|
" <td>53943.000000</td>\n",
|
|||
|
|
" <td>53943.000000</td>\n",
|
|||
|
|
" <td>53943.000000</td>\n",
|
|||
|
|
" <td>53943.000000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>mean</th>\n",
|
|||
|
|
" <td>26972.000000</td>\n",
|
|||
|
|
" <td>0.797935</td>\n",
|
|||
|
|
" <td>61.749322</td>\n",
|
|||
|
|
" <td>57.457251</td>\n",
|
|||
|
|
" <td>3932.734294</td>\n",
|
|||
|
|
" <td>5.731158</td>\n",
|
|||
|
|
" <td>5.734526</td>\n",
|
|||
|
|
" <td>3.538730</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>std</th>\n",
|
|||
|
|
" <td>15572.147122</td>\n",
|
|||
|
|
" <td>0.473999</td>\n",
|
|||
|
|
" <td>1.432626</td>\n",
|
|||
|
|
" <td>2.234549</td>\n",
|
|||
|
|
" <td>3989.338447</td>\n",
|
|||
|
|
" <td>1.121730</td>\n",
|
|||
|
|
" <td>1.142103</td>\n",
|
|||
|
|
" <td>0.705679</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>min</th>\n",
|
|||
|
|
" <td>1.000000</td>\n",
|
|||
|
|
" <td>0.200000</td>\n",
|
|||
|
|
" <td>43.000000</td>\n",
|
|||
|
|
" <td>43.000000</td>\n",
|
|||
|
|
" <td>326.000000</td>\n",
|
|||
|
|
" <td>0.000000</td>\n",
|
|||
|
|
" <td>0.000000</td>\n",
|
|||
|
|
" <td>0.000000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>25%</th>\n",
|
|||
|
|
" <td>13486.500000</td>\n",
|
|||
|
|
" <td>0.400000</td>\n",
|
|||
|
|
" <td>61.000000</td>\n",
|
|||
|
|
" <td>56.000000</td>\n",
|
|||
|
|
" <td>950.000000</td>\n",
|
|||
|
|
" <td>4.710000</td>\n",
|
|||
|
|
" <td>4.720000</td>\n",
|
|||
|
|
" <td>2.910000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>50%</th>\n",
|
|||
|
|
" <td>26972.000000</td>\n",
|
|||
|
|
" <td>0.700000</td>\n",
|
|||
|
|
" <td>61.800000</td>\n",
|
|||
|
|
" <td>57.000000</td>\n",
|
|||
|
|
" <td>2401.000000</td>\n",
|
|||
|
|
" <td>5.700000</td>\n",
|
|||
|
|
" <td>5.710000</td>\n",
|
|||
|
|
" <td>3.530000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>75%</th>\n",
|
|||
|
|
" <td>40457.500000</td>\n",
|
|||
|
|
" <td>1.040000</td>\n",
|
|||
|
|
" <td>62.500000</td>\n",
|
|||
|
|
" <td>59.000000</td>\n",
|
|||
|
|
" <td>5324.000000</td>\n",
|
|||
|
|
" <td>6.540000</td>\n",
|
|||
|
|
" <td>6.540000</td>\n",
|
|||
|
|
" <td>4.040000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>max</th>\n",
|
|||
|
|
" <td>53943.000000</td>\n",
|
|||
|
|
" <td>5.010000</td>\n",
|
|||
|
|
" <td>79.000000</td>\n",
|
|||
|
|
" <td>95.000000</td>\n",
|
|||
|
|
" <td>18823.000000</td>\n",
|
|||
|
|
" <td>10.740000</td>\n",
|
|||
|
|
" <td>58.900000</td>\n",
|
|||
|
|
" <td>31.800000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" </tbody>\n",
|
|||
|
|
"</table>\n",
|
|||
|
|
"</div>"
|
|||
|
|
],
|
|||
|
|
"text/plain": [
|
|||
|
|
" Unnamed: 0 carat depth table price \\\n",
|
|||
|
|
"count 53943.000000 53943.000000 53943.000000 53943.000000 53943.000000 \n",
|
|||
|
|
"mean 26972.000000 0.797935 61.749322 57.457251 3932.734294 \n",
|
|||
|
|
"std 15572.147122 0.473999 1.432626 2.234549 3989.338447 \n",
|
|||
|
|
"min 1.000000 0.200000 43.000000 43.000000 326.000000 \n",
|
|||
|
|
"25% 13486.500000 0.400000 61.000000 56.000000 950.000000 \n",
|
|||
|
|
"50% 26972.000000 0.700000 61.800000 57.000000 2401.000000 \n",
|
|||
|
|
"75% 40457.500000 1.040000 62.500000 59.000000 5324.000000 \n",
|
|||
|
|
"max 53943.000000 5.010000 79.000000 95.000000 18823.000000 \n",
|
|||
|
|
"\n",
|
|||
|
|
" x y z \n",
|
|||
|
|
"count 53943.000000 53943.000000 53943.000000 \n",
|
|||
|
|
"mean 5.731158 5.734526 3.538730 \n",
|
|||
|
|
"std 1.121730 1.142103 0.705679 \n",
|
|||
|
|
"min 0.000000 0.000000 0.000000 \n",
|
|||
|
|
"25% 4.710000 4.720000 2.910000 \n",
|
|||
|
|
"50% 5.700000 5.710000 3.530000 \n",
|
|||
|
|
"75% 6.540000 6.540000 4.040000 \n",
|
|||
|
|
"max 10.740000 58.900000 31.800000 "
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"execution_count": 18,
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "execute_result"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Описание данных (основные статистические показатели)\n",
|
|||
|
|
"df.describe()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 19,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"text/plain": [
|
|||
|
|
"Unnamed: 0 0\n",
|
|||
|
|
"carat 0\n",
|
|||
|
|
"cut 0\n",
|
|||
|
|
"color 0\n",
|
|||
|
|
"clarity 0\n",
|
|||
|
|
"depth 0\n",
|
|||
|
|
"table 0\n",
|
|||
|
|
"price 0\n",
|
|||
|
|
"x 0\n",
|
|||
|
|
"y 0\n",
|
|||
|
|
"z 0\n",
|
|||
|
|
"dtype: int64"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"execution_count": 19,
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "execute_result"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Процент пропущенных значений признаков\n",
|
|||
|
|
"for i in df.columns:\n",
|
|||
|
|
" null_rate = df[i].isnull().sum() / len(df) * 100\n",
|
|||
|
|
" if null_rate > 0:\n",
|
|||
|
|
" print(f'{i} Процент пустых значений: %{null_rate:.2f}')\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Проверка на пропущенные данные\n",
|
|||
|
|
"df.isnull().sum()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 20,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"text/plain": [
|
|||
|
|
"Unnamed: 0 False\n",
|
|||
|
|
"carat False\n",
|
|||
|
|
"cut False\n",
|
|||
|
|
"color False\n",
|
|||
|
|
"clarity False\n",
|
|||
|
|
"depth False\n",
|
|||
|
|
"table False\n",
|
|||
|
|
"price False\n",
|
|||
|
|
"x False\n",
|
|||
|
|
"y False\n",
|
|||
|
|
"z False\n",
|
|||
|
|
"dtype: bool"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"execution_count": 20,
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "execute_result"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"df.isnull().any()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"Пропущенных данных в колонках нет :)"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"# Анализ сведений о наборе данных\n",
|
|||
|
|
"\n",
|
|||
|
|
"**Набор данных:** Набор данных содержит 53 940 наблюдений, каждое из которых представляет собой бриллиант с характеристиками, такими как вес (карат), огранка, цвет, прозрачность, глубина, размеры (x, y, z), а также цена.\n",
|
|||
|
|
"\n",
|
|||
|
|
"**Проблемная область:** Проблемной областью может быть задача оценки стоимости бриллианта на основе его характеристик, что важно для торговли бриллиантами, определения их рыночной стоимости и других бизнес-задач.\n",
|
|||
|
|
"\n",
|
|||
|
|
"Проанализируем содержимое...\n",
|
|||
|
|
"\n",
|
|||
|
|
"*Объекты наблюдения:* Каждый бриллиант будет объектом наблюдения. \n",
|
|||
|
|
"\n",
|
|||
|
|
"*Атрибуты объектов:* Основные атрибуты представляют собой числовые характеристики такие, как carat (вес), depth (глубина в бриллианте - расстояние от площадки до каллеты. Калетта - нижняя часть бриллианта), table (площадка - большая грань бриллианта), x,y,z - размеры. Категориальные характеристики такие, как cut (огранка), color (цвет) и clarity (прозрачность). \n",
|
|||
|
|
"\n",
|
|||
|
|
"*Связи между объектами:* Вес бриллианта влияет на цену (carat -> price), также существует связь между качеством огранки и ценой (cut -> price)\n",
|
|||
|
|
"\n",
|
|||
|
|
"## Бизнес-цель\n",
|
|||
|
|
"1. Автоматизация процесса оценки стоимости бриллианта \n",
|
|||
|
|
"Построить модель для оценки цены бриллианты на основе его характеристик. Эффект для бизнеса от анализа данных заключается в улучшении бизнес-процессов (повышение эффективности операционных процессов), повышении доходов, сокращении затрат и улучшению клиентских отношений. \n",
|
|||
|
|
"2. Оптимизация ассортимента и продаж \n",
|
|||
|
|
"Построить модель для анализа данных, какие характеристики бриллиантов наиболее востребованы и наилучшим образом влияют на рост продаж. Эффект для бизнеса в том, что компании могут оптимизировать ассортимент на более востребованные, что приведёт к увеличению продаж.\n",
|
|||
|
|
"\n",
|
|||
|
|
"### Техническая цель\n",
|
|||
|
|
"1. Разработка модели для оценки стоимости бриллиантов \n",
|
|||
|
|
"Создать модель машинного обучения, которая на вход получает характеристики бриллианта (вес, огранка, цвет и т.д.) и предсказывает его цену. \n",
|
|||
|
|
"На вход поступают характеристики, а целевой признак - цена бриллианта. \n",
|
|||
|
|
"2. Оптимизация ассортимента и прогнозирование спроса \n",
|
|||
|
|
"Создать модель для предсказания спроса (популярности) определённых категорий бриллиаантов в зависимости от их характеристик. НА вход поступают характеристики бриллианта, а целевой признак - показатели продаж или востребованность товара."
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 21,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 2000x1500 with 4 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Построим графики boxplot для обнаружения выбросов\n",
|
|||
|
|
"plt.figure(figsize=(20, 15))\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Создание boxplot\n",
|
|||
|
|
"for i, column in enumerate(['carat', 'depth', 'table', 'price'], 1):\n",
|
|||
|
|
" plt.subplot(4, 2, i)\n",
|
|||
|
|
" sns.boxplot(x=df[column])\n",
|
|||
|
|
" plt.title(f\"Boxplot для {column}\")\n",
|
|||
|
|
" \n",
|
|||
|
|
"plt.tight_layout()\n",
|
|||
|
|
"plt.show()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"## Проблемы наборов\n",
|
|||
|
|
"*Зашумленность данных:* Может быть обнаружена через большое количество выбросов (аномальных значений). Выбросы могут выходить за нормальные границы. Они наблюдаются во всех графиках, которые мы построили. Если у определённого параметра много выбросов, то данные могут быть зашумлены или содержат необычные значения, которые стоит анализировать и, возможно, удалить или скорректировать. Графики boxplot помогают визуализировать распределение данных и выявить выбросы, которые находятся за пределами \"усов\" боксплота.\n",
|
|||
|
|
"\n",
|
|||
|
|
"Попробуем решить устранить проблему выбросов для table и depth. Используется метод усреднения данных для устранения выбросов. Метод стандартных отклонений."
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 22,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 2000x1500 with 3 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Вычисление среднего и стандартного отклонения для столбцов\n",
|
|||
|
|
"means = df[['carat', 'depth', 'table']].mean()\n",
|
|||
|
|
"std_devs = df[['carat', 'depth', 'table']].std()\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Определение порога для выбросов (например, 3 стандартных отклонения)\n",
|
|||
|
|
"threshold = 3\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Функция для замены выбросов на среднее значение\n",
|
|||
|
|
"def replace_outliers_with_mean(column):\n",
|
|||
|
|
" mean_value = column.mean()\n",
|
|||
|
|
" upper_limit = mean_value + threshold * column.std()\n",
|
|||
|
|
" lower_limit = mean_value - threshold * column.std()\n",
|
|||
|
|
" \n",
|
|||
|
|
" return column.where((column >= lower_limit) & (column <= upper_limit), mean_value)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Применение функции к каждому столбцу\n",
|
|||
|
|
"for column in ['depth', 'table', 'carat']:\n",
|
|||
|
|
" df[column] = replace_outliers_with_mean(df[column])\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Графическое отображение после замены выбросов\n",
|
|||
|
|
"plt.figure(figsize=(20, 15))\n",
|
|||
|
|
"for i, column in enumerate(['carat', 'depth', 'table'], 1):\n",
|
|||
|
|
" plt.subplot(4, 2, i)\n",
|
|||
|
|
" sns.boxplot(x=df[column])\n",
|
|||
|
|
" plt.title(f\"Boxplot для {column} после замены выбросов на среднее значение\")\n",
|
|||
|
|
"\n",
|
|||
|
|
"plt.tight_layout()\n",
|
|||
|
|
"plt.show()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"Этот подход позволяет вам обрабатывать выбросы, не прибегая к квантилям и он наверяка избавит от выбросов. Однако стоит отметить, что использование стандартных отклонений может не всегда быть оптимальным для всех распределений данных, потому что данные могут быть плохо распределены. А использование медианы с квантилями вместо среднего значения помогает лучше справиться с выбросами, особенно если распределение данных имеет сильные отклонения."
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 23,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 2000x1500 with 3 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Вычисление квантилей для каждого столбца\n",
|
|||
|
|
"Q1 = df[['carat', 'depth', 'table']].quantile(0.01)\n",
|
|||
|
|
"Q3 = df[['carat', 'depth', 'table']].quantile(0.99)\n",
|
|||
|
|
"IQR = Q3 - Q1\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Определение границ для выбросов\n",
|
|||
|
|
"lower_bound = Q1 - 1.5 * IQR\n",
|
|||
|
|
"upper_bound = Q3 + 1.5 * IQR\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Функция для замены выбросов на медианное значение\n",
|
|||
|
|
"def replace_outliers_with_median(column):\n",
|
|||
|
|
" median_value = column.median()\n",
|
|||
|
|
" return column.where((column >= lower_bound[column.name]) & (column <= upper_bound[column.name]), median_value)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Применение функции к каждому столбцу\n",
|
|||
|
|
"for column in ['depth', 'table', 'carat']:\n",
|
|||
|
|
" df[column] = replace_outliers_with_median(df[column])\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Графическое отображение после замены выбросов\n",
|
|||
|
|
"plt.figure(figsize=(20, 15))\n",
|
|||
|
|
"for i, column in enumerate(['carat', 'depth', 'table'], 1):\n",
|
|||
|
|
" plt.subplot(4, 2, i)\n",
|
|||
|
|
" sns.boxplot(x=df[column])\n",
|
|||
|
|
" plt.title(f\"Boxplot для {column} после замены выбросов на медиану\")\n",
|
|||
|
|
"\n",
|
|||
|
|
"plt.tight_layout()\n",
|
|||
|
|
"plt.show()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"*Актуальность данных* — это проблема, которая возникает, когда данные устарели или больше не отражают текущую ситуацию. Это часто встречается в данных о ценах. \n",
|
|||
|
|
"*Просачивание данных* происходит, когда в обучающей выборке присутствует информация, которая не должна быть известна модели на этапе обучения. В данном случае это может происходить, если некоторые переменные, такие как цена, сильно коррелируют с целевым признаком и тем самым искажают результат. Проверка на наличие утечек данных. \n",
|
|||
|
|
"*Смещение в данных* может возникать, если датасет не полностью репрезентативен для реальной популяции бриллиантов. Например, если в датасете больше дешёвых или дорогих бриллиантов, это создаёт смещение в оценке моделей \n",
|
|||
|
|
"Код ниже проверяет данные на распредение признаков и просачивание данных."
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 24,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1500x1000 with 2 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1000x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Проверка распределения признаков\n",
|
|||
|
|
"plt.figure(figsize=(15, 10))\n",
|
|||
|
|
"for i, column in enumerate(['price', 'cut'], 1):\n",
|
|||
|
|
" plt.subplot(3, 1, i)\n",
|
|||
|
|
" sns.histplot(df[column], bins=20, kde=True)\n",
|
|||
|
|
" plt.title(f\"Распределение {column}\")\n",
|
|||
|
|
"plt.tight_layout()\n",
|
|||
|
|
"plt.show()\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Средняя цена по типу огранки (cut)\n",
|
|||
|
|
"cut_price_mean = df.groupby('cut')['price'].mean().sort_values()\n",
|
|||
|
|
"plt.figure(figsize=(10, 6))\n",
|
|||
|
|
"sns.barplot(x=cut_price_mean.index, y=cut_price_mean.values)\n",
|
|||
|
|
"plt.title(\"Средняя цена по типу огранки (cut)\")\n",
|
|||
|
|
"plt.xlabel(\"Огранка\")\n",
|
|||
|
|
"plt.ylabel(\"Средняя цена\")\n",
|
|||
|
|
"plt.show()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"Для решения проблем с просачиваниям, проведём корелляционый анализ между признаками и целевым признаком. "
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 3,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"name": "stdout",
|
|||
|
|
"output_type": "stream",
|
|||
|
|
"text": [
|
|||
|
|
"Корреляция признаков с ценой (price):\n",
|
|||
|
|
"price 1.000000\n",
|
|||
|
|
"carat 0.921591\n",
|
|||
|
|
"x 0.884433\n",
|
|||
|
|
"y 0.865419\n",
|
|||
|
|
"z 0.861249\n",
|
|||
|
|
"clarity_SI2 0.128427\n",
|
|||
|
|
"table 0.127118\n",
|
|||
|
|
"color_I 0.097130\n",
|
|||
|
|
"cut_Premium 0.095685\n",
|
|||
|
|
"color_J 0.081714\n",
|
|||
|
|
"color_H 0.059229\n",
|
|||
|
|
"clarity_SI1 0.008940\n",
|
|||
|
|
"color_G 0.008564\n",
|
|||
|
|
"cut_Very Good 0.006589\n",
|
|||
|
|
"cut_Good -0.000307\n",
|
|||
|
|
"clarity_VS2 -0.001066\n",
|
|||
|
|
"clarity_VS1 -0.009879\n",
|
|||
|
|
"depth -0.010630\n",
|
|||
|
|
"color_F -0.024166\n",
|
|||
|
|
"clarity_IF -0.049593\n",
|
|||
|
|
"clarity_VVS2 -0.052375\n",
|
|||
|
|
"clarity_VVS1 -0.095261\n",
|
|||
|
|
"cut_Ideal -0.097160\n",
|
|||
|
|
"color_E -0.101101\n",
|
|||
|
|
"Unnamed: 0 -0.306875\n",
|
|||
|
|
"Name: price, dtype: float64\n"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1000x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"from sklearn.model_selection import train_test_split\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Преобразуем все категориальные данные в числовые\n",
|
|||
|
|
"# Проверим, какие столбцы являются категориальными\n",
|
|||
|
|
"categorical_columns = df.select_dtypes(include=['object']).columns\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Применяем one-hot encoding к категориальным столбцам\n",
|
|||
|
|
"df_encoded = pd.get_dummies(df, columns=categorical_columns, drop_first=True)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Вычисляем корреляцию для преобразованных данных\n",
|
|||
|
|
"correlations = df_encoded.corr()['price'].sort_values(ascending=False)\n",
|
|||
|
|
"print(\"Корреляция признаков с ценой (price):\")\n",
|
|||
|
|
"print(correlations)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Визуализация распределения цены (price)\n",
|
|||
|
|
"plt.figure(figsize=(10, 6))\n",
|
|||
|
|
"sns.histplot(df['price'], bins=30, kde=True)\n",
|
|||
|
|
"plt.title('Распределение цены (price)')\n",
|
|||
|
|
"plt.xlabel('Цена')\n",
|
|||
|
|
"plt.ylabel('Количество')\n",
|
|||
|
|
"plt.show()\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Разделение данных на обучающую и тестовую выборки\n",
|
|||
|
|
"X = df_encoded.drop('price', axis=1)\n",
|
|||
|
|
"y = df_encoded['price']\n",
|
|||
|
|
"\n",
|
|||
|
|
"X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42, stratify=df['cut'])\n"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 26,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"name": "stdout",
|
|||
|
|
"output_type": "stream",
|
|||
|
|
"text": [
|
|||
|
|
"Количество дубликатов: 0\n"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "iVBORw0KGgoAAAANSUhEUgAAA/8AAAIjCAYAAABViau2AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAACFfklEQVR4nOzdd3xT9f7H8XdauqEtpXQgqwzZBSmKFUWUQhkucCIqIIIiqIAicq8C6vWH4lUBQdCrggMUHLgQEIqASEEZlY2ARUBo2S0UOmjP74+Q0HTRkTZt8no+HrlJTr4555OQ5vrOdxyTYRiGAAAAAACA03JzdAEAAAAAAKB8Ef4BAAAAAHByhH8AAAAAAJwc4R8AAAAAACdH+AcAAAAAwMkR/gEAAAAAcHKEfwAAAAAAnBzhHwAAAAAAJ0f4BwDACZ05c0b79+9XWlqao0sBnJphGDp58qT27Nnj6FIAoEiEfwAAnIBhGHrvvfd07bXXytfXV/7+/oqIiNCnn37q6NKqhEOHDmnOnDnW+/v379fcuXMdVxC0bds2ffPNN9b7CQkJWrRokeMKyuXMmTN6/vnn1axZM3l6eqpWrVq68sortXv3bkeXBgCFquboAgDAEebMmaNBgwZZ73t5eal+/frq3r27XnjhBYWGhjqwOqDk7r//fs2fP18DBgzQmDFjFBAQIJPJpMjISEeXViWYTCYNHz5c4eHhatasmZ599lkFBQWpf//+ji7NZZ05c0aPPvqowsLCVKtWLT311FPq2bOnevfu7dC6Tpw4oRtvvFEHDhzQE088oU6dOsnT01MeHh5q2LChQ2sDgKIQ/gG4tJdeekkRERFKT0/XmjVrNHPmTP3444/atm2bfH19HV0eUCwff/yx5s+fr08//VT333+/o8upkq644goNGTJEPXr0kCSFh4dr5cqVji3KxUVHR1svknTllVdqyJAhDq5KGjNmjI4cOaL4+Hi1atXK0eUAQLGZDMMwHF0EAFQ0S8//77//rg4dOli3P/3003rzzTc1b9489evXz4EVAsXXpk0bRUZGMkzdDvbt26fjx4+rdevW8vPzc3Q5kLRjxw6dP39ebdq0kaenp0NrOXr0qMLDwzVr1qxK8UMEAJQEc/4BIJebb75ZkpSYmChJOnnypJ555hm1adNG1atXl7+/v3r27Kk//vgj33PT09M1ceJEXXnllfL29lZ4eLj69u2rffv2STLPITaZTIVeunTpYt3XypUrZTKZNH/+fP3rX/9SWFiY/Pz8dNttt+ngwYP5jr1+/Xr16NFDAQEB8vX11Y033qhff/21wNfYpUuXAo8/ceLEfG0//fRTRUVFycfHR0FBQbrvvvsKPH5Rry23nJwcTZkyRa1atZK3t7dCQ0P16KOP6tSpUzbtGjZsqFtuuSXfcUaMGJFvnwXV/vrrr+d7TyUpIyNDEyZMUJMmTeTl5aV69erp2WefVUZGRoHvlcWcOXOK9e9WkLzvd3BwsHr37q1t27YV+TyLL774wvpvEBwcrAceeED//POP9fG0tDRt27ZN9erVU+/eveXv7y8/Pz916dJFv/zyi7XdX3/9JZPJpLfeeivfMdauXSuTyaTPPvvMWnPe12X5N849L37Lli0aOHCgGjVqJG9vb4WFhenhhx/WiRMnCnz/9u/fb922dOlSXXfddfL19VVAQIBuueWWfO/JxIkTZTKZdPz4ceu2DRs25KtDklq3bl3gv8XixYt1ww03yM/PTzVq1FDv3r21fft2mzYDBw60Dtdu3LixOnbsqJMnT8rHxydf3QXJ/XyLgwcPFuv5AwcOLPJ7Ie/zi/N6JGnXrl265557VLt2bfn4+KhZs2b697//LenS+1rUJfeoh8t9Bgt6HTVr1sz3GbR455131KpVK3l5ealOnToaPny4Tp8+bdMm92ewZcuWioqK0h9//FHg90pBivt3ZzKZNGLEiEL3k/ez+/vvvysnJ0eZmZnq0KGDvL29VatWLfXr108HDhzI9/wVK1ZY/70CAwN1++23a+fOnTZtLP8eln8zf39/6zSH9PT0fPXm/r67cOGCevXqpaCgIO3YscO6ffbs2br55psVEhIiLy8vtWzZUjNnzsxXm5ubm8aPH2+zfd68eTKZTPnaA6j6GPYPALlYgnqtWrUkmQPTN998o7vvvlsRERFKTk7Wu+++qxtvvFE7duxQnTp1JEnZ2dm65ZZbFBcXp/vuu09PPfWUzpw5o2XLlmnbtm1q3Lix9Rj9+vVTr169bI47bty4Aut55ZVXZDKZNHbsWB09elRTpkxRTEyMEhIS5OPjI8n8H3A9e/ZUVFSUJkyYIDc3N+t/+P3yyy+65ppr8u23bt26mjRpkiTp7NmzGjZsWIHHfuGFF3TPPffokUce0bFjx/T222+rc+fO2rx5swIDA/M9Z+jQobrhhhskSV9//bUWLlxo8/ijjz5qHXXx5JNPKjExUdOnT9fmzZv166+/ysPDo8D3oSROnz5tfW255eTk6LbbbtOaNWs0dOhQtWjRQlu3btVbb72lP//802Zhsbw6d+6sTz75xGbb33//reeff14hISGXral58+b697//LcMwtG/fPr355pvq1atXgWEhN8t7dfXVV2vSpElKTk7W1KlT9euvv1r/DSxB+7XXXlNYWJjGjBkjb29v/e9//1NMTIyWLVumzp07q1GjRurUqZPmzp2rUaNG2Rxn7ty5qlGjhm6//fbLvpbcli1bpr/++kuDBg1SWFiYtm/frvfee0/bt2/XunXrCg1pv/zyi3r16qUGDRpowoQJysrK0jvvvKNOnTrp999/15VXXlmiOgrzySefaMCAAYqNjdVrr72mc+fOaebMmbr++uu1efPmIudnjx8/Pl/wKoniPv/RRx9VTEyM9f6DDz6oPn36qG/fvtZttWvXLtHr2bJli2644QZ5eHho6NChatiwofbt26fvv/9er7zyivr27asmTZpY9z9q1Ci1aNFCQ4cOtW5r0aKFpOJ9Bi2Cg4OtPy4dOnRIU6dOVa9evXTw4EFru4kTJ+rFF19UTEyMhg0bpt27d2vmzJn6/fffL/sdMHbs2Mu+n7mV9u+uKJa/txEjRigqKkqvvvqqjh07pmnTpmnNmjXavHmzgoODJUnLly9Xz5491ahRI02cOFHnz5/X22+/rU6dOmnTpk35Pn/33HOPGjZsqEmTJmndunWaNm2aTp06pY8//rjQeh555BGtXLlSy5YtU8uWLa3bZ86cqVatWum2225TtWrV9P333+vxxx9XTk6Ohg8fLsn8Y/fjjz+uSZMm6Y477lD79u115MgRPfHEE4qJidFjjz1W6vcJQCVlAIALmj17tiHJWL58uXHs2DHj4MGDxueff27UqlXL8PHxMQ4dOmQYhmGkp6cb2dnZNs9NTEw0vLy8jJdeesm67cMPPzQkGW+++Wa+Y+Xk5FifJ8l4/fXX87Vp1aqVceONN1rv//zzz4Yk44orrjBSU1Ot2xcsWGBIMqZOnWrdd9OmTY3Y2FjrcQzDMM6dO2dEREQY3bp1y3es6667zmjdurX1/rFjxwxJxoQJE6zb9u/fb7i7uxuvvPKKzXO3bt1qVKtWLd/2PXv2GJKMjz76yLptwoQJRu7/m/nll18MScbcuXNtnrtkyZJ82xs0aGD07t07X+3Dhw838v5fV97an332WSMkJMSIioqyeU8/+eQTw83Nzfjll19snj9r1ixDkvHrr7/mO15hzp8/b0RFRRl16tQxjhw5UmTbG2+80aYOwzCMf/3rX4Yk4+jRo4U+LzMz0wgJCTFat25tnD9/3rr9hx9+MCQZ48ePNwzj0ufK09PT+PPPP63tjh07ZtSqVcuIioqybnv33XcNScbOnTttjhMcHGwMGDDAuu2mm24yOnfubFOP5TizZ8+2bjt37ly+uj/77DNDkrF69WrrNsvfW2JiomEYhhEVFWUEBAQYSUlJ1jZ//vmn4eHhYdx5553WbZbP0LFjx6zbfv/993x1GEb+v6EzZ84YgYGBxpAhQ2zaJSUlGQEBATbbBwwYYDRo0MB6f9u2bYabm5vRs2dPm7oLU9b
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "iVBORw0KGgoAAAANSUhEUgAAA/8AAAIjCAYAAABViau2AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAAB80klEQVR4nOzde3yP9f/H8edn5/NmttkwzPnMzxxCIcYcSk7JoRwiJZIU8q0cUolOhOhIBxKFinI+JYdqyCGEaMQwzGzsfP3++LRP+9gcNts+8/G4327XbZ/Pdb2v63pdH9c+9rwO78tkGIYhAAAAAABgtxxsXQAAAAAAAChYhH8AAAAAAOwc4R8AAAAAADtH+AcAAAAAwM4R/gEAAAAAsHOEfwAAAAAA7BzhHwAAAAAAO0f4BwAAAADAzhH+AQBAgUhMTNTx48d14cIFW5cC2L24uDgdPnxYaWlpti4FQBFF+AcAoADMnTtXJpNJx44du63XkVuLFi1Sq1at5O3tLS8vL5UpU0ZTpkyxdVm3hYSEBE2dOtXyPi4uTjNnzrRdQdCJEyc0d+5cy/tjx45p3rx5tisoi9TUVE2ZMkV16tSRq6urihUrpkqVKmnt2rW2Lg1AEUX4B1DoMgNL5uDm5qbKlStr6NChOn36tK3Lg527fPmyxo8frw0bNuR63mPHjslkMunNN9/M/8LswPPPP6/u3bvL29tbH374oVavXq01a9boySeftHVptwV3d3e9+OKLmjdvno4fP67x48fr+++/t3VZdzSTyaQhQ4Zo5cqVOnbsmEaNGqWffvrJ1mUpOTlZEREReumll9SiRQstWrRIq1ev1rp169S4cWNblwegiHKydQEA7lwvv/yywsLClJSUpM2bN2vWrFn64YcftHfvXnl4eNi6PNipy5cva8KECZKkFi1a2LYYO7Jx40ZNnjxZkyZN0vPPP2/rcm5Ljo6OmjBhgvr06aOMjAz5+Pho+fLlti7rjlaqVCk99thjatu2rSQpJCQkTwcO89vkyZO1fft2rVy5ku8xADeN8A/AZtq1a6f69etLkgYOHKjixYvr7bff1rfffquePXvauDoAufHmm2+qSZMmBP9b9Oyzz+qhhx7S8ePHVa1aNfn5+dm6pDve1KlT9dRTTyk2NlY1a9aUp6enTetJS0vT1KlT9eyzzxL8AeQKl/0DKDJatmwpSTp69Kgk6fz583ruuedUq1YteXl5ycfHR+3atdPvv/+ebd6kpCSNHz9elStXlpubm0JCQtSlSxcdOXJE0n+Xa19ryPoH1IYNG2QymfTVV1/pf//7n4KDg+Xp6amOHTvq+PHj2da9fft2tW3bVr6+vvLw8FDz5s31888/57iNLVq0yHH948ePz9b2iy++UHh4uNzd3eXv768ePXrkuP7rbVtWGRkZmjp1qmrUqCE3NzeVKFFCjz/+eLbO2MqVK6f77rsv23qGDh2abZk51f7GG29k+0wl82Wq48aNU8WKFeXq6qrQ0FCNGjVKycnJOX5WV9u+fbvat2+vYsWKydPTU7Vr19a0adMs01u0aJHjH8L9+vVTuXLlJJk/q8DAQEnShAkTrvv558a+ffvUsmVLubu7q3Tp0nrllVeUkZGRY9sff/xR99xzjzw9PeXt7a0OHTpo3759Vm12796tfv36qXz58nJzc1NwcLAeffRRnTt3Lk/19evXz2q/KFasmFq0aHHTly+vW7fOUrOfn58eeOAB7d+/36rNtm3bVLNmTfXo0UP+/v5yd3dXgwYNtHTpUkubhIQEeXp66umnn862jhMnTsjR0VGTJk2y1Jz575bV1f9ef//9t5588klVqVJF7u7uKl68uB588MFs/SBk/l5nPWv766+/qnXr1vL29panp2eOn0nmbUq//fabZVxsbGyO+819992XY8038x0xfvx4y+9X6dKl1bhxYzk5OSk4ODhb3TnJOn+mhISEm5o/c97rDVnnv9nvvH/++UcDBgxQyZIl5erqqrCwMA0ePFgpKSnZbv/Kach6r/3N7INXb4e3t7caNmxotQ9mWrRokeX7NSAgQA8//LD++ecfqzZZ98EKFSqoUaNGOn/+vNzd3W+qr42b/b271ndupqv33YMHD+rChQvy9vZW8+bN5eHhIV9fX913333au3dvtvl37typdu3aycfHR15eXmrVqpW2bdtm1Sbz32PTpk16/PHHVbx4cfn4+KhPnz45/h/Rr18/q3GDBg2Sm5ub1X7y7bffqkOHDpZ//woVKmjixIlKT0+3tNm/f7/c3d3Vp08fq+Vt3rxZjo6OGj169DU/FwC5x5l/AEVGZlAvXry4JOmvv/7S0qVL9eCDDyosLEynT5/W+++/r+bNm+uPP/5QyZIlJUnp6em67777tHbtWvXo0UNPP/20Ll26pNWrV2vv3r2qUKGCZR09e/ZU+/btrdY7ZsyYHOt59dVXZTKZNHr0aJ05c0ZTp05VRESEdu3aJXd3d0nmP0jbtWun8PBwjRs3Tg4ODpozZ45atmypn376SQ0bNsy23NKlS1sCTkJCggYPHpzjul966SV1795dAwcO1NmzZzV9+nQ1a9ZMO3fuzPFs4KBBg3TPPfdIkhYvXqwlS5ZYTX/88cc1d+5c9e/fX8OGDdPRo0c1Y8YM7dy5Uz///LOcnZ1z/BxyIy4uzrJtWWVkZKhjx47avHmzBg0apGrVqmnPnj1655139Oeff+b4x3lWq1ev1n333aeQkBA9/fTTCg4O1v79+7Vs2bIcg+S1BAYGatasWRo8eLA6d+6sLl26SJJq166dq+3MKiYmRvfee6/S0tL0/PPPy9PTUx988IFlH8nq888/V9++fRUZGanJkyfr8uXLmjVrlu6++27t3LnTEjRWr16tv/76S/3791dwcLD27dunDz74QPv27dO2bduyhbybERAQoHfeeUeSOWhPmzZN7du31/Hjx697dnnNmjVq166dypcvr/Hjx+vKlSuaPn26mjZtqh07dlhqPnfunD744AN5eXlp2LBhCgwM1BdffKEuXbpo3rx56tmzp7y8vNS5c2d99dVXevvtt+Xo6GhZz5dffinDMNS7d+9cbdevv/6qLVu2qEePHipdurSOHTumWbNmqUWLFvrjjz+ueQvR4cOH1aJFC3l4eGjkyJHy8PDQhx9+qIiICK1evVrNmjXLVR3XkpfviExvvfXWLfWDcrPzd+nSRRUrVrS8f+aZZ1StWjUNGjTIMq5atWq52p6TJ0+qYcOGiouL06BBg1S1alX9888/+vrrr3X58mU1a9ZMn3/+uWX5r776qiTphRdesIxr0qSJpJvfBzNlLjc2NlbvvfeeHnzwQe3du1dVqlSRJMv3YIMGDTRp0iSdPn1a06ZN088//3zN79dMY8eOVVJS0g0/00x5/b27nsyDgGPGjFGlSpU0YcIEJSUlaebMmWratKl+/fVXVa5cWZL5wOQ999wjHx8fjRo1Ss7Oznr//ffVokULbdy4UY0aNbJa9tChQ+Xn56fx48fr4MGDmjVrlv7++2/LAYicjBs3Th9//LG++uorqwOwc+fOlZeXl0aMGCEvLy+tW7dOY8eOVXx8vN544w1J5v1q4sSJGjlypLp166aOHTsqMTFR/fr1U9WqVfXyyy/n6TMCcA0GABSyOXPmGJKMNWvWGGfPnjWOHz9uLFiwwChevLjh7u5unDhxwjAMw0hKSjLS09Ot5j169Kjh6upqvPzyy5Zxn3zyiSHJePvtt7OtKyMjwzKfJOONN97I1qZGjRpG8+bNLe/Xr19vSDJKlSplxMfHW8YvXLjQkGRMmzbNsuxKlSoZkZGRlvUYhmFcvnzZCAsLM1q3bp1tXU2aNDFq1qxpeX/27FlDkjFu3DjLuGPHjhmOjo7Gq6++ajXvnj17DCcnp2zjDx06ZEgyPv30U8u4cePGGVm/4n/66SdDkjFv3jyreVesWJFtfNmyZY0OHTpkq33IkCHG1f9tXF37qFGjjKCgICM8PNzqM/38888NBwcH46effrKaf/bs2YYk4+eff862vkxpaWlGWFiYUbZsWePChQtW07J+7s2bN7daZ6a+ffs
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "iVBORw0KGgoAAAANSUhEUgAAA/8AAAIjCAYAAABViau2AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAAB+zUlEQVR4nOzde3yP9f/H8edn59nRzGwY5nwmEy2FGHOIUIkkp1BRoVK+FdJBqUQRP510oByKDjQxhw5GGXKWmEY2zGGbYcfr98enffKxYZttn+3jcb/drts+n+t6X9f1uj67Pp/t+bne13WZDMMwBAAAAAAA7JaDrQsAAAAAAADFi/APAAAAAICdI/wDAAAAAGDnCP8AAAAAANg5wj8AAAAAAHaO8A8AAAAAgJ0j/AMAAAAAYOcI/wAAAAAA2DnCPwAAKJSUlBQdPnxYqampti4FsGuGYej06dM6cOCArUsBUIYR/gEAuMT8+fNlMpl0+PBhW5dS6hiGoXnz5umWW25RuXLl5O3trZCQEH3++ee2Lq1MOHr0qObPn295fvjwYS1YsMB2BUG7du3S8uXLLc+3b9+uFStW2K6gS6SkpOj5559XvXr15OLiogoVKqhu3brav3+/rUsDUEYR/gGUOjnhK2dwc3NT3bp1NXr0aB0/ftzW5QHF4vL9/tLh2WeftXV5kqT7779fDz/8sBo0aKDPPvtMq1ev1po1a9SnTx9bl1YmmEwmjRo1SqtWrdLhw4c1fvx4/fzzz7Yu64aWkpKikSNHatOmTTpw4ICeeOIJ7dy509Zl6dSpUwoLC9M777yje+65R998841Wr16t9evXq0aNGrYuD0AZ5WTrAgDgSqZMmaKQkBBdvHhRv/zyi+bMmaOVK1dq165dKleunK3LA4pFzn5/qcaNG9uomv98+umnWrRokT7//HPdf//9ti6nTKpSpYqGDx+uLl26SJKCgoK0fv162xZ1gwsLC7MMklS3bl0NHz7cxlVJTz/9tOLj4xUdHa1GjRrZuhwAdoLwD6DU6tq1q1q2bClJeuihh1ShQgVNnz5d33zzjfr372/j6oD8OX/+fIG+rLp0vy9N3njjDfXv35/gf51mzJihxx57TImJiWrcuLE8PDxsXdINb/ny5dqzZ48uXLigJk2ayMXFxab1nDhxQp988onmzp1L8AdQpOj2D6DM6NChgyQpNjZWknT69Gk99dRTatKkiTw9PeXt7a2uXbvqjz/+yDXvxYsXNXnyZNWtW1dubm4KCgpSnz59dPDgQUnmc2+v1OXaZDKpffv2lmWtX79eJpNJixYt0v/+9z8FBgbKw8NDPXv21JEjR3Kte/PmzerSpYt8fHxUrlw5tWvXTr/++mue29i+ffs81z958uRcbT///HOFhobK3d1dfn5+6tevX57rv9q2XSo7O1szZsxQo0aN5ObmpkqVKmnkyJE6c+aMVbsaNWrozjvvzLWe0aNH51pmXrW/8cYbuV5TSUpLS9OkSZNUu3Ztubq6Kjg4WOPHj1daWlqer9XlNm/erG7duql8+fLy8PBQ06ZNNXPmTKs2a9eu1e233y4PDw/5+vrqrrvu0t69e/O1/Pfee0+NGjWSq6urKleurFGjRuns2bNWbdq3b6/GjRsrJiZGbdu2Vbly5fS///0vX8u/HpfvN/7+/urevbt27dqVr/mXLFli2Zf8/f31wAMP6J9//rFMT01N1a5duxQcHKzu3bvL29tbHh4eat++vVW39UOHDslkMuntt9/OtY6NGzfKZDLpiy++sNR8+T6Qs69eel78jh07NHjwYNWsWVNubm4KDAzU0KFDderUKat587pWw6pVq3TrrbeqXLly8vHx0Z133pnrNZk8ebJMJpMSExMt47Zs2ZKrDsncA+PymiXphx9+sOxXXl5e6t69u3bv3m3VZvDgwZbu2rVq1VLr1q11+vRpubu75+saE5fOn+PIkSP5mn/w4MFX/Xy7fP78bI8k7du3T3379lXFihXl7u6uevXq6bnnnpP03+t6teHSXg/X2gfz2o7y5cvn2gdz5Pf9mvP7bNiwoUJDQ/XHH3/k+fmYl/y+70wmk0aPHn3F5Vy+7/7+++/Kzs5Wenq6WrZsKTc3N1WoUEH9+/dXXFxcrvnz87mW8/vI+Z15e3urQoUKeuKJJ3Tx4sVc9V76uZ2Zmalu3brJz89Pe/bssYz/+OOP1aFDBwUEBMjV1VUNGzbUnDlzctXm4OCgiRMnWo1fuHChTCZTrvYAihdH/gGUGTlBvUKFCpLMQWP58uW69957FRISouPHj+v//u//1K5dO+3Zs0eVK1eWJGVlZenOO+9UVFSU+vXrpyeeeEIpKSlavXq1du3apVq1alnW0b9/f3Xr1s1qvRMmTMiznldeeUUmk0nPPPOMTpw4oRkzZig8PFzbt2+Xu7u7JPM/Pl27dlVoaKgmTZokBwcHyz9MP//8s1q1apVruVWrVtXUqVMlSefOndMjjzyS57pfeOEF9e3bVw899JBOnjypd999V23bttW2bdvk6+uba54RI0bo9ttvlyR9/fXXWrZsmdX0kSNHav78+RoyZIgef/xxxcbGatasWdq2bZt+/fVXOTs75/k6FMTZs2ct23ap7Oxs9ezZU7/88otGjBihBg0aaOfOnXr77bf1559/Wl2QKy+rV6/WnXfeqaCgID3xxBMKDAzU3r179f333+uJJ56QJK1Zs0Zdu3ZVzZo1NXnyZF24cEHvvvuu2rRpo61bt171PNrJkyfrxRdfVHh4uB555BHt379fc+bM0e+//57rtTl16pS6du2qfv366YEHHlClSpUK9BolJSVZBVFJ8vf3v+Z89evX13PPPSfDMHTw4EFNnz5d3bp1yzMsXCrnd37zzTdr6tSpOn78uGbOnKlff/3Vsi/lBO3XX39dgYGBevrpp+Xm5qb3339f4eHhWr16tdq2bauaNWuqTZs2WrBggcaOHWu1ngULFsjLy0t33XVXgV6P1atX69ChQxoyZIgCAwO1e/duzZs3T7t379amTZuuGNJ+/vlndevWTdWrV9ekSZOUkZGh9957T23atNHvv/+uunXrFqiOK/nss880aNAgRURE6PXXX9f58+c1Z84c3Xbbbdq2bdtV96uJEyfmCl4Fkd/5R44cqfDwcMvzgQMHqnfv3lbXaqhYsWKBtmfHjh26/fbb5ezsrBEjRqhGjRo6ePCgvvvuO73yyivq06ePateubVn+2LFj1aBBA40YMcIyrkGDBpLytw/m8Pf3t3y5dPToUc2cOVPdunXTkSNHLO0K8n693DPPPHPN1/NShX3fXU3O+2306NEKDQ3Va6+9ppMnT+qdd97RL7/8om3btlk+Ewr6uda3b1/VqFFDU6dO1aZNm/TOO+/ozJkz+vTTT69Yz0MPPaT169dr9erVatiwoWX8nDlz1KhRI/Xs2VNOTk767rvv9Oijjyo7O1ujRo2SZP7S/tFHH9XUqVPVq1cvtWjRQvHx8XrssccUHh6uhx9+uNCvE4BCMACglPn4448NScaaNWuMkydPGkeOHDG+/PJLo0KFCoa7u7tx9OhRwzAM4+LFi0ZWVpbVvLGxsYarq6sxZcoUy7iPPvrIkGRMnz4917qys7Mt80ky3njjjVxtGjVqZLRr187yfN26dYYko0qVKkZycrJl/OLFiw1JxsyZMy3LrlOnjhEREWFZj2EYxvnz542QkBCjU6dOudZ16623Go0bN7Y8P3nypCHJmDRpkmXc4cOHDUdHR+OVV16xmnfnzp2Gk5NTrvEHDhwwJBmffPKJZdykSZOMS/8E/Pzzz4YkY8GCBVbzRkZG5hpfvXp1o3v37rlqHzVqlHH5n5XLax8/frwREBBghIaGWr2mn332meHg4GD8/PPPVvPPnTvXkGT8+uuvudaXIzMz0wgJCTGqV69unDlzxmrapa978+bNjYCAAOPUqVOWcX/88Yfh4OBgPPjgg5ZxOftfbGysYRiGceLECcPFxcXo3Lmz1f42a9YsQ5Lx0UcfWca1a9fOkGTMnTv3ivVeSc568xqupV27dlavp2EYxv/+9z9DknHixIkrzpe
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"name": "stdout",
|
|||
|
|
"output_type": "stream",
|
|||
|
|
"text": [
|
|||
|
|
"Корреляция признаков в обучающей выборке:\n",
|
|||
|
|
" Unnamed: 0 carat depth table x y \\\n",
|
|||
|
|
"Unnamed: 0 1.000000 -0.397135 -0.038173 -0.109717 -0.406396 -0.404400 \n",
|
|||
|
|
"carat -0.397135 1.000000 0.012244 0.184243 0.951931 0.945835 \n",
|
|||
|
|
"depth -0.038173 0.012244 1.000000 -0.274986 -0.030762 -0.032418 \n",
|
|||
|
|
"table -0.109717 0.184243 -0.274986 1.000000 0.195246 0.188413 \n",
|
|||
|
|
"x -0.406396 0.951931 -0.030762 0.195246 1.000000 0.991915 \n",
|
|||
|
|
"y -0.404400 0.945835 -0.032418 0.188413 0.991915 1.000000 \n",
|
|||
|
|
"z -0.399257 0.928342 0.076986 0.152752 0.967789 0.962323 \n",
|
|||
|
|
"cut_Good -0.042108 0.034264 0.192864 0.163059 0.027477 0.030305 \n",
|
|||
|
|
"cut_Ideal 0.114386 -0.163612 -0.020022 -0.560157 -0.162808 -0.155536 \n",
|
|||
|
|
"cut_Premium -0.052857 0.111731 -0.225003 0.373514 0.125573 0.107317 \n",
|
|||
|
|
"cut_Very Good -0.035069 0.017264 0.034832 0.137452 0.007004 0.019726 \n",
|
|||
|
|
"color_E 0.052187 -0.132396 -0.023732 0.007044 -0.132446 -0.130469 \n",
|
|||
|
|
"color_F 0.016246 -0.048969 -0.016200 -0.007303 -0.046116 -0.045376 \n",
|
|||
|
|
"color_G 0.027986 -0.024513 -0.003837 -0.038404 -0.026801 -0.026822 \n",
|
|||
|
|
"color_H -0.062449 0.100708 0.022672 0.017285 0.095782 0.093429 \n",
|
|||
|
|
"color_I -0.045077 0.153477 0.021523 0.014445 0.149141 0.148534 \n",
|
|||
|
|
"color_J -0.050254 0.152898 0.021672 0.040373 0.162742 0.162404 \n",
|
|||
|
|
"clarity_IF 0.077743 -0.119400 -0.033011 -0.078131 -0.130391 -0.127557 \n",
|
|||
|
|
"clarity_SI1 -0.053072 0.073320 0.050914 0.055756 0.080001 0.078622 \n",
|
|||
|
|
"clarity_SI2 -0.160840 0.255131 0.006453 0.097077 0.272073 0.267797 \n",
|
|||
|
|
"clarity_VS1 0.025491 -0.056952 -0.022840 -0.030577 -0.058078 -0.055083 \n",
|
|||
|
|
"clarity_VS2 0.032349 -0.033837 -0.006613 -0.010324 -0.036775 -0.037425 \n",
|
|||
|
|
"clarity_VVS1 0.095239 -0.171270 -0.020592 -0.069129 -0.188238 -0.185525 \n",
|
|||
|
|
"clarity_VVS2 0.091095 -0.131283 -0.017795 -0.062135 -0.143086 -0.140215 \n",
|
|||
|
|
"\n",
|
|||
|
|
" z cut_Good cut_Ideal cut_Premium ... color_H \\\n",
|
|||
|
|
"Unnamed: 0 -0.399257 -0.042108 0.114386 -0.052857 ... -0.062449 \n",
|
|||
|
|
"carat 0.928342 0.034264 -0.163612 0.111731 ... 0.100708 \n",
|
|||
|
|
"depth 0.076986 0.192864 -0.020022 -0.225003 ... 0.022672 \n",
|
|||
|
|
"table 0.152752 0.163059 -0.560157 0.373514 ... 0.017285 \n",
|
|||
|
|
"x 0.967789 0.027477 -0.162808 0.125573 ... 0.095782 \n",
|
|||
|
|
"y 0.962323 0.030305 -0.155536 0.107317 ... 0.093429 \n",
|
|||
|
|
"z 1.000000 0.042419 -0.158216 0.087557 ... 0.094510 \n",
|
|||
|
|
"cut_Good 0.042419 1.000000 -0.258011 -0.185394 ... -0.007642 \n",
|
|||
|
|
"cut_Ideal -0.158216 -0.258011 1.000000 -0.478081 ... -0.024151 \n",
|
|||
|
|
"cut_Premium 0.087557 -0.185394 -0.478081 1.000000 ... 0.029929 \n",
|
|||
|
|
"cut_Very Good 0.018725 -0.169940 -0.438229 -0.314889 ... -0.005621 \n",
|
|||
|
|
"color_E -0.129694 0.007180 0.001097 -0.017061 ... -0.201856 \n",
|
|||
|
|
"color_F -0.046366 0.005680 0.002736 -0.014361 ... -0.197792 \n",
|
|||
|
|
"color_G -0.026922 -0.024765 0.035195 0.002625 ... -0.220750 \n",
|
|||
|
|
"color_H 0.094510 -0.007642 -0.024151 0.029929 ... 1.000000 \n",
|
|||
|
|
"color_I 0.146920 0.006280 -0.011736 0.007090 ... -0.142441 \n",
|
|||
|
|
"color_J 0.161595 0.016478 -0.039704 0.019348 ... -0.100320 \n",
|
|||
|
|
"clarity_IF -0.129263 -0.032327 0.104625 -0.053332 ... 0.008773 \n",
|
|||
|
|
"clarity_SI1 0.081263 0.056069 -0.081767 0.023309 ... 0.034336 \n",
|
|||
|
|
"clarity_SI2 0.263557 0.043577 -0.108704 0.065610 ... 0.024856 \n",
|
|||
|
|
"clarity_VS1 -0.055805 -0.020710 0.034568 -0.013400 ... -0.015491 \n",
|
|||
|
|
"clarity_VS2 -0.037401 -0.021250 0.014164 0.024069 ... -0.033375 \n",
|
|||
|
|
"clarity_VVS1 -0.184765 -0.035799 0.088938 -0.052519 ... 0.001434 \n",
|
|||
|
|
"clarity_VVS2 -0.140629 -0.037519 0.075342 -0.060646 ... -0.030919 \n",
|
|||
|
|
"\n",
|
|||
|
|
" color_I color_J clarity_IF clarity_SI1 clarity_SI2 \\\n",
|
|||
|
|
"Unnamed: 0 -0.045077 -0.050254 0.077743 -0.053072 -0.160840 \n",
|
|||
|
|
"carat 0.153477 0.152898 -0.119400 0.073320 0.255131 \n",
|
|||
|
|
"depth 0.021523 0.021672 -0.033011 0.050914 0.006453 \n",
|
|||
|
|
"table 0.014445 0.040373 -0.078131 0.055756 0.097077 \n",
|
|||
|
|
"x 0.149141 0.162742 -0.130391 0.080001 0.272073 \n",
|
|||
|
|
"y 0.148534 0.162404 -0.127557 0.078622 0.267797 \n",
|
|||
|
|
"z 0.146920 0.161595 -0.129263 0.081263 0.263557 \n",
|
|||
|
|
"cut_Good 0.006280 0.016478 -0.032327 0.056069 0.043577 \n",
|
|||
|
|
"cut_Ideal -0.011736 -0.039704 0.104625 -0.081767 -0.108704 \n",
|
|||
|
|
"cut_Premium 0.007090 0.019348 -0.053332 0.023309 0.065610 \n",
|
|||
|
|
"cut_Very Good 0.000094 0.008815 -0.033901 0.031762 0.005171 \n",
|
|||
|
|
"color_E -0.157077 -0.110628 -0.043930 0.000156 0.005365 \n",
|
|||
|
|
"color_F -0.153915 -0.108401 0.015369 -0.017832 -0.003781 \n",
|
|||
|
|
"color_G -0.171780 -0.120983 0.078387 -0.083663 -0.045632 \n",
|
|||
|
|
"color_H -0.142441 -0.100320 0.008773 0.034336 0.024856 \n",
|
|||
|
|
"color_I 1.000000 -0.078066 -0.013824 0.019822 -0.001150 \n",
|
|||
|
|
"color_J -0.078066 1.000000 -0.018986 0.016113 -0.000187 \n",
|
|||
|
|
"clarity_IF -0.013824 -0.018986 1.000000 -0.105528 -0.084748 \n",
|
|||
|
|
"clarity_SI1 0.019822 0.016113 -0.105528 1.000000 -0.256125 \n",
|
|||
|
|
"clarity_SI2 -0.001150 -0.000187 -0.084748 -0.256125 1.000000 \n",
|
|||
|
|
"clarity_VS1 0.019454 0.024517 -0.078583 -0.237493 -0.190726 \n",
|
|||
|
|
"clarity_VS2 -0.007408 0.018553 -0.101376 -0.306378 -0.246046 \n",
|
|||
|
|
"clarity_VVS1 -0.003091 -0.038902 -0.050966 -0.154031 -0.123699 \n",
|
|||
|
|
"clarity_VVS2 -0.032959 -0.036901 -0.059965 -0.181225 -0.145538 \n",
|
|||
|
|
"\n",
|
|||
|
|
" clarity_VS1 clarity_VS2 clarity_VVS1 clarity_VVS2 \n",
|
|||
|
|
"Unnamed: 0 0.025491 0.032349 0.095239 0.091095 \n",
|
|||
|
|
"carat -0.056952 -0.033837 -0.171270 -0.131283 \n",
|
|||
|
|
"depth -0.022840 -0.006613 -0.020592 -0.017795 \n",
|
|||
|
|
"table -0.030577 -0.010324 -0.069129 -0.062135 \n",
|
|||
|
|
"x -0.058078 -0.036775 -0.188238 -0.143086 \n",
|
|||
|
|
"y -0.055083 -0.037425 -0.185525 -0.140215 \n",
|
|||
|
|
"z -0.055805 -0.037401 -0.184765 -0.140629 \n",
|
|||
|
|
"cut_Good -0.020710 -0.021250 -0.035799 -0.037519 \n",
|
|||
|
|
"cut_Ideal 0.034568 0.014164 0.088938 0.075342 \n",
|
|||
|
|
"cut_Premium -0.013400 0.024069 -0.052519 -0.060646 \n",
|
|||
|
|
"cut_Very Good -0.003510 -0.015913 -0.008381 0.014051 \n",
|
|||
|
|
"color_E -0.023478 0.031569 -0.004026 0.011237 \n",
|
|||
|
|
"color_F -0.012930 0.004648 0.018415 0.012658 \n",
|
|||
|
|
"color_G 0.059925 -0.028513 0.046097 0.061750 \n",
|
|||
|
|
"color_H -0.015491 -0.033375 0.001434 -0.030919 \n",
|
|||
|
|
"color_I 0.019454 -0.007408 -0.003091 -0.032959 \n",
|
|||
|
|
"color_J 0.024517 0.018553 -0.038902 -0.036901 \n",
|
|||
|
|
"clarity_IF -0.078583 -0.101376 -0.050966 -0.059965 \n",
|
|||
|
|
"clarity_SI1 -0.237493 -0.306378 -0.154031 -0.181225 \n",
|
|||
|
|
"clarity_SI2 -0.190726 -0.246046 -0.123699 -0.145538 \n",
|
|||
|
|
"clarity_VS1 1.000000 -0.228148 -0.114701 -0.134951 \n",
|
|||
|
|
"clarity_VS2 -0.228148 1.000000 -0.147970 -0.174094 \n",
|
|||
|
|
"clarity_VVS1 -0.114701 -0.147970 1.000000 -0.087525 \n",
|
|||
|
|
"clarity_VVS2 -0.134951 -0.174094 -0.087525 1.000000 \n",
|
|||
|
|
"\n",
|
|||
|
|
"[24 rows x 24 columns]\n"
|
|||
|
|
]
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Проверка на наличие дубликатов\n",
|
|||
|
|
"duplicates = df.duplicated().sum()\n",
|
|||
|
|
"print(f\"Количество дубликатов: {duplicates}\")\n",
|
|||
|
|
"\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Проверим, какие столбцы являются категориальными\n",
|
|||
|
|
"categorical_columns = X.select_dtypes(include=['object']).columns\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Применяем one-hot encoding к категориальным столбцам\n",
|
|||
|
|
"X_encoded = pd.get_dummies(X, columns=categorical_columns, drop_first=True)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Разделение на обучающую и тестовую выборки\n",
|
|||
|
|
"X_train, X_test, y_train, y_test = train_test_split(X_encoded, y, test_size=0.2, random_state=42, stratify=df['cut'])\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Визуализация распределений признаков\n",
|
|||
|
|
"def visualize_distribution(X_train, X_test, feature):\n",
|
|||
|
|
" plt.figure(figsize=(12, 6))\n",
|
|||
|
|
" sns.histplot(X_train[feature], bins=30, color='blue', label='Train', kde=True)\n",
|
|||
|
|
" sns.histplot(X_test[feature], bins=30, color='red', label='Test', kde=True)\n",
|
|||
|
|
" plt.title(f'Распределение {feature} в обучающей и тестовой выборках')\n",
|
|||
|
|
" plt.xlabel(feature)\n",
|
|||
|
|
" plt.ylabel('Количество')\n",
|
|||
|
|
" plt.legend()\n",
|
|||
|
|
" plt.show()\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Распределение для каждого из признаков\n",
|
|||
|
|
"for feature in X_encoded.columns:\n",
|
|||
|
|
" visualize_distribution(X_train, X_test, feature)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Проверка корреляции\n",
|
|||
|
|
"correlations = X_train.corr()\n",
|
|||
|
|
"print(\"Корреляция признаков в обучающей выборке:\")\n",
|
|||
|
|
"print(correlations)\n",
|
|||
|
|
"\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Преобразование категориальных данных в числовые\n",
|
|||
|
|
"df_encoded = pd.get_dummies(df, columns=['cut'], drop_first=True)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Разделение данных на обучающую и тестовую выборки\n",
|
|||
|
|
"X = df_encoded.drop('price', axis=1)\n",
|
|||
|
|
"y = df_encoded['price']\n",
|
|||
|
|
"\n",
|
|||
|
|
"X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"## Оценка выборок и Аугментация данных"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 27,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"name": "stdout",
|
|||
|
|
"output_type": "stream",
|
|||
|
|
"text": [
|
|||
|
|
"Сбалансированность в обучающей выборке:\n",
|
|||
|
|
"price\n",
|
|||
|
|
"605 109\n",
|
|||
|
|
"789 106\n",
|
|||
|
|
"828 102\n",
|
|||
|
|
"625 101\n",
|
|||
|
|
"544 98\n",
|
|||
|
|
" ... \n",
|
|||
|
|
"8952 1\n",
|
|||
|
|
"13769 1\n",
|
|||
|
|
"10719 1\n",
|
|||
|
|
"6064 1\n",
|
|||
|
|
"3169 1\n",
|
|||
|
|
"Name: count, Length: 10680, dtype: int64\n",
|
|||
|
|
"\n",
|
|||
|
|
"Сбалансированность в тестовой выборке:\n",
|
|||
|
|
"price\n",
|
|||
|
|
"802 33\n",
|
|||
|
|
"776 29\n",
|
|||
|
|
"844 29\n",
|
|||
|
|
"675 26\n",
|
|||
|
|
"645 25\n",
|
|||
|
|
" ..\n",
|
|||
|
|
"7976 1\n",
|
|||
|
|
"7022 1\n",
|
|||
|
|
"8969 1\n",
|
|||
|
|
"5074 1\n",
|
|||
|
|
"4968 1\n",
|
|||
|
|
"Name: count, Length: 5338, dtype: int64\n"
|
|||
|
|
]
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Проверка сбалансированности целевой переменной в обучающей выборке\n",
|
|||
|
|
"cut_counts_train = y_train.value_counts()\n",
|
|||
|
|
"cut_counts_test = y_test.value_counts()\n",
|
|||
|
|
"\n",
|
|||
|
|
"print(\"Сбалансированность в обучающей выборке:\")\n",
|
|||
|
|
"print(cut_counts_train)\n",
|
|||
|
|
"print(\"\\nСбалансированность в тестовой выборке:\")\n",
|
|||
|
|
"print(cut_counts_test)"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"После того как мы оценили сбалансированность, можем решить, нужно ли использовать методы аугментации данных. Например, если одна из категорий в 'cut' значительно меньше других, это может потребовать применения oversampling или undersampling."
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 4,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"name": "stdout",
|
|||
|
|
"output_type": "stream",
|
|||
|
|
"text": [
|
|||
|
|
"Количество дубликатов: 0\n",
|
|||
|
|
"Распределение классов в обучающей выборке:\n",
|
|||
|
|
"price\n",
|
|||
|
|
"605 106\n",
|
|||
|
|
"802 105\n",
|
|||
|
|
"789 103\n",
|
|||
|
|
"544 99\n",
|
|||
|
|
"828 99\n",
|
|||
|
|
" ... \n",
|
|||
|
|
"12690 1\n",
|
|||
|
|
"12304 1\n",
|
|||
|
|
"8972 1\n",
|
|||
|
|
"18018 1\n",
|
|||
|
|
"5634 1\n",
|
|||
|
|
"Name: count, Length: 10695, dtype: int64\n",
|
|||
|
|
"\n",
|
|||
|
|
"Сбалансированность после oversampling:\n",
|
|||
|
|
"price\n",
|
|||
|
|
"12592 106\n",
|
|||
|
|
"6750 106\n",
|
|||
|
|
"3750 106\n",
|
|||
|
|
"506 106\n",
|
|||
|
|
"2583 106\n",
|
|||
|
|
" ... \n",
|
|||
|
|
"1189 106\n",
|
|||
|
|
"1013 106\n",
|
|||
|
|
"956 106\n",
|
|||
|
|
"4637 106\n",
|
|||
|
|
"3601 106\n",
|
|||
|
|
"Name: count, Length: 6622, dtype: int64\n"
|
|||
|
|
]
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Проверка на наличие дубликатов\n",
|
|||
|
|
"duplicates = df.duplicated().sum()\n",
|
|||
|
|
"print(f\"Количество дубликатов: {duplicates}\")\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Преобразуем все категориальные данные в числовые\n",
|
|||
|
|
"categorical_columns = df.select_dtypes(include=['object']).columns\n",
|
|||
|
|
"df_encoded = pd.get_dummies(df, columns=categorical_columns, drop_first=True)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Разделение данных на обучающую и тестовую выборки\n",
|
|||
|
|
"X = df_encoded.drop('price', axis=1)\n",
|
|||
|
|
"y = df_encoded['price']\n",
|
|||
|
|
"\n",
|
|||
|
|
"X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42, stratify=df['cut'])\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Проверка сбалансированности целевой переменной\n",
|
|||
|
|
"class_counts = y_train.value_counts()\n",
|
|||
|
|
"print(\"Распределение классов в обучающей выборке:\")\n",
|
|||
|
|
"print(class_counts)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Фильтрация классов с недостаточным количеством образцов\n",
|
|||
|
|
"# Установим минимальное количество образцов, чтобы сохранить класс\n",
|
|||
|
|
"min_samples = 2\n",
|
|||
|
|
"classes_to_keep = class_counts[class_counts >= min_samples].index\n",
|
|||
|
|
"\n",
|
|||
|
|
"X_train_filtered = X_train[y_train.isin(classes_to_keep)]\n",
|
|||
|
|
"y_train_filtered = y_train[y_train.isin(classes_to_keep)]\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Применение SMOTE или RandomOverSampler\n",
|
|||
|
|
"from imblearn.over_sampling import SMOTE, RandomOverSampler\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Используем SMOTE, если классов достаточно\n",
|
|||
|
|
"if y_train_filtered.value_counts().min() > 1:\n",
|
|||
|
|
" smote = SMOTE(random_state=42, k_neighbors=1)\n",
|
|||
|
|
" X_resampled, y_resampled = smote.fit_resample(X_train_filtered, y_train_filtered)\n",
|
|||
|
|
"else:\n",
|
|||
|
|
" # Или RandomOverSampler\n",
|
|||
|
|
" ros = RandomOverSampler(random_state=42)\n",
|
|||
|
|
" X_resampled, y_resampled = ros.fit_resample(X_train_filtered, y_train_filtered)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Проверка сбалансированности после oversampling\n",
|
|||
|
|
"resampled_counts = y_resampled.value_counts()\n",
|
|||
|
|
"print(\"\\nСбалансированность после oversampling:\")\n",
|
|||
|
|
"print(resampled_counts)"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"Итог произвели балансировку данных. Использовались методы oversampling (увеличение данных недопредставленных категорий) и\n",
|
|||
|
|
"методы SMOTE. Произвелась аугментация данных (создавались дополительные данные из имеющихся). Использование методов, таких как SMOTE или RandomOverSampler, помогло увеличить количество экземпляров редких классов в обучающей выборке, что повысило шансы модели правильно классифицировать их. \n",
|
|||
|
|
"\n",
|
|||
|
|
"\n",
|
|||
|
|
"### 2 Датасет: Цены на ноутбуки \n",
|
|||
|
|
"Выгрузили второй датасет! (цены ноутбуки)"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 43,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"name": "stdout",
|
|||
|
|
"output_type": "stream",
|
|||
|
|
"text": [
|
|||
|
|
"Index(['Company', 'Product', 'TypeName', 'Inches', 'Ram', 'OS', 'Weight',\n",
|
|||
|
|
" 'Price_euros', 'Screen', 'ScreenW', 'ScreenH', 'Touchscreen',\n",
|
|||
|
|
" 'IPSpanel', 'RetinaDisplay', 'CPU_company', 'CPU_freq', 'CPU_model',\n",
|
|||
|
|
" 'PrimaryStorage', 'SecondaryStorage', 'PrimaryStorageType',\n",
|
|||
|
|
" 'SecondaryStorageType', 'GPU_company', 'GPU_model'],\n",
|
|||
|
|
" dtype='object')\n",
|
|||
|
|
"\n",
|
|||
|
|
"<class 'pandas.core.frame.DataFrame'>\n",
|
|||
|
|
"RangeIndex: 1275 entries, 0 to 1274\n",
|
|||
|
|
"Data columns (total 23 columns):\n",
|
|||
|
|
" # Column Non-Null Count Dtype \n",
|
|||
|
|
"--- ------ -------------- ----- \n",
|
|||
|
|
" 0 Company 1275 non-null object \n",
|
|||
|
|
" 1 Product 1275 non-null object \n",
|
|||
|
|
" 2 TypeName 1275 non-null object \n",
|
|||
|
|
" 3 Inches 1275 non-null float64\n",
|
|||
|
|
" 4 Ram 1275 non-null int64 \n",
|
|||
|
|
" 5 OS 1275 non-null object \n",
|
|||
|
|
" 6 Weight 1275 non-null float64\n",
|
|||
|
|
" 7 Price_euros 1275 non-null float64\n",
|
|||
|
|
" 8 Screen 1275 non-null object \n",
|
|||
|
|
" 9 ScreenW 1275 non-null int64 \n",
|
|||
|
|
" 10 ScreenH 1275 non-null int64 \n",
|
|||
|
|
" 11 Touchscreen 1275 non-null object \n",
|
|||
|
|
" 12 IPSpanel 1275 non-null object \n",
|
|||
|
|
" 13 RetinaDisplay 1275 non-null object \n",
|
|||
|
|
" 14 CPU_company 1275 non-null object \n",
|
|||
|
|
" 15 CPU_freq 1275 non-null float64\n",
|
|||
|
|
" 16 CPU_model 1275 non-null object \n",
|
|||
|
|
" 17 PrimaryStorage 1275 non-null int64 \n",
|
|||
|
|
" 18 SecondaryStorage 1275 non-null int64 \n",
|
|||
|
|
" 19 PrimaryStorageType 1275 non-null object \n",
|
|||
|
|
" 20 SecondaryStorageType 1275 non-null object \n",
|
|||
|
|
" 21 GPU_company 1275 non-null object \n",
|
|||
|
|
" 22 GPU_model 1275 non-null object \n",
|
|||
|
|
"dtypes: float64(4), int64(5), object(14)\n",
|
|||
|
|
"memory usage: 229.2+ KB\n"
|
|||
|
|
]
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"df = pd.read_csv(\".//static//csv//laptop_prices.csv\")\n",
|
|||
|
|
"print(df.columns)\n",
|
|||
|
|
"\n",
|
|||
|
|
"print()\n",
|
|||
|
|
"\n",
|
|||
|
|
"df.info()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"*Датасет:* Цена ноутбука. Прайс-лист компании-производителя ноутбуков для регрессии \n",
|
|||
|
|
"\n",
|
|||
|
|
"Kaggle предоставляет следующие данные о датасете: Компания — производитель ноутбуков, Продукт — бренд и модель, Название типа — тип (ноутбук,ультрабук, игровой и т. д.), Дюймы — размер экрана, Разрешение экрана — разрешение экрана, Процессор — центральный процессор (ЦП), Оперативная память — оперативная память ноутбука, Память — жёсткий диск / SSD-накопитель, GPU — графические процессоры (GPU), OpSys — операционная система, Вес — вес ноутбука, Цена_в_евро — цена (в евро)\n",
|
|||
|
|
"\n",
|
|||
|
|
"*Объекты наблюдения:* Каждый объект наблюдения в наборе данных — это конкретная модель ноутбука, представленная в прайс-листе производителя.\n",
|
|||
|
|
"*Атрибуты объектов:* компания, продукт, тип, диагональ экрана ОЗУ (оперативная память), Операционная система, вес, цена, разрешение экрана, IPS-панель, модель GPU, модель процессора (CPU_model), частота процессора, производитель CPU.\n",
|
|||
|
|
"\n",
|
|||
|
|
"*Связи между объектами:* Взаимосвязь между характеристиками ноутбуков и их ценой. Например, цена может зависеть от типа процессора, объема памяти, разрешения экрана, а также от производителя."
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 6,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"text/html": [
|
|||
|
|
"<div>\n",
|
|||
|
|
"<style scoped>\n",
|
|||
|
|
" .dataframe tbody tr th:only-of-type {\n",
|
|||
|
|
" vertical-align: middle;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"\n",
|
|||
|
|
" .dataframe tbody tr th {\n",
|
|||
|
|
" vertical-align: top;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"\n",
|
|||
|
|
" .dataframe thead th {\n",
|
|||
|
|
" text-align: right;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"</style>\n",
|
|||
|
|
"<table border=\"1\" class=\"dataframe\">\n",
|
|||
|
|
" <thead>\n",
|
|||
|
|
" <tr style=\"text-align: right;\">\n",
|
|||
|
|
" <th></th>\n",
|
|||
|
|
" <th>Company</th>\n",
|
|||
|
|
" <th>Product</th>\n",
|
|||
|
|
" <th>TypeName</th>\n",
|
|||
|
|
" <th>Inches</th>\n",
|
|||
|
|
" <th>Ram</th>\n",
|
|||
|
|
" <th>OS</th>\n",
|
|||
|
|
" <th>Weight</th>\n",
|
|||
|
|
" <th>Price_euros</th>\n",
|
|||
|
|
" <th>Screen</th>\n",
|
|||
|
|
" <th>ScreenW</th>\n",
|
|||
|
|
" <th>...</th>\n",
|
|||
|
|
" <th>RetinaDisplay</th>\n",
|
|||
|
|
" <th>CPU_company</th>\n",
|
|||
|
|
" <th>CPU_freq</th>\n",
|
|||
|
|
" <th>CPU_model</th>\n",
|
|||
|
|
" <th>PrimaryStorage</th>\n",
|
|||
|
|
" <th>SecondaryStorage</th>\n",
|
|||
|
|
" <th>PrimaryStorageType</th>\n",
|
|||
|
|
" <th>SecondaryStorageType</th>\n",
|
|||
|
|
" <th>GPU_company</th>\n",
|
|||
|
|
" <th>GPU_model</th>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" </thead>\n",
|
|||
|
|
" <tbody>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>0</th>\n",
|
|||
|
|
" <td>Apple</td>\n",
|
|||
|
|
" <td>MacBook Pro</td>\n",
|
|||
|
|
" <td>Ultrabook</td>\n",
|
|||
|
|
" <td>13.3</td>\n",
|
|||
|
|
" <td>8</td>\n",
|
|||
|
|
" <td>macOS</td>\n",
|
|||
|
|
" <td>1.37</td>\n",
|
|||
|
|
" <td>1339.69</td>\n",
|
|||
|
|
" <td>Standard</td>\n",
|
|||
|
|
" <td>2560</td>\n",
|
|||
|
|
" <td>...</td>\n",
|
|||
|
|
" <td>Yes</td>\n",
|
|||
|
|
" <td>Intel</td>\n",
|
|||
|
|
" <td>2.3</td>\n",
|
|||
|
|
" <td>Core i5</td>\n",
|
|||
|
|
" <td>128</td>\n",
|
|||
|
|
" <td>0</td>\n",
|
|||
|
|
" <td>SSD</td>\n",
|
|||
|
|
" <td>No</td>\n",
|
|||
|
|
" <td>Intel</td>\n",
|
|||
|
|
" <td>Iris Plus Graphics 640</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>1</th>\n",
|
|||
|
|
" <td>Apple</td>\n",
|
|||
|
|
" <td>Macbook Air</td>\n",
|
|||
|
|
" <td>Ultrabook</td>\n",
|
|||
|
|
" <td>13.3</td>\n",
|
|||
|
|
" <td>8</td>\n",
|
|||
|
|
" <td>macOS</td>\n",
|
|||
|
|
" <td>1.34</td>\n",
|
|||
|
|
" <td>898.94</td>\n",
|
|||
|
|
" <td>Standard</td>\n",
|
|||
|
|
" <td>1440</td>\n",
|
|||
|
|
" <td>...</td>\n",
|
|||
|
|
" <td>No</td>\n",
|
|||
|
|
" <td>Intel</td>\n",
|
|||
|
|
" <td>1.8</td>\n",
|
|||
|
|
" <td>Core i5</td>\n",
|
|||
|
|
" <td>128</td>\n",
|
|||
|
|
" <td>0</td>\n",
|
|||
|
|
" <td>Flash Storage</td>\n",
|
|||
|
|
" <td>No</td>\n",
|
|||
|
|
" <td>Intel</td>\n",
|
|||
|
|
" <td>HD Graphics 6000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>2</th>\n",
|
|||
|
|
" <td>HP</td>\n",
|
|||
|
|
" <td>250 G6</td>\n",
|
|||
|
|
" <td>Notebook</td>\n",
|
|||
|
|
" <td>15.6</td>\n",
|
|||
|
|
" <td>8</td>\n",
|
|||
|
|
" <td>No OS</td>\n",
|
|||
|
|
" <td>1.86</td>\n",
|
|||
|
|
" <td>575.00</td>\n",
|
|||
|
|
" <td>Full HD</td>\n",
|
|||
|
|
" <td>1920</td>\n",
|
|||
|
|
" <td>...</td>\n",
|
|||
|
|
" <td>No</td>\n",
|
|||
|
|
" <td>Intel</td>\n",
|
|||
|
|
" <td>2.5</td>\n",
|
|||
|
|
" <td>Core i5 7200U</td>\n",
|
|||
|
|
" <td>256</td>\n",
|
|||
|
|
" <td>0</td>\n",
|
|||
|
|
" <td>SSD</td>\n",
|
|||
|
|
" <td>No</td>\n",
|
|||
|
|
" <td>Intel</td>\n",
|
|||
|
|
" <td>HD Graphics 620</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>3</th>\n",
|
|||
|
|
" <td>Apple</td>\n",
|
|||
|
|
" <td>MacBook Pro</td>\n",
|
|||
|
|
" <td>Ultrabook</td>\n",
|
|||
|
|
" <td>15.4</td>\n",
|
|||
|
|
" <td>16</td>\n",
|
|||
|
|
" <td>macOS</td>\n",
|
|||
|
|
" <td>1.83</td>\n",
|
|||
|
|
" <td>2537.45</td>\n",
|
|||
|
|
" <td>Standard</td>\n",
|
|||
|
|
" <td>2880</td>\n",
|
|||
|
|
" <td>...</td>\n",
|
|||
|
|
" <td>Yes</td>\n",
|
|||
|
|
" <td>Intel</td>\n",
|
|||
|
|
" <td>2.7</td>\n",
|
|||
|
|
" <td>Core i7</td>\n",
|
|||
|
|
" <td>512</td>\n",
|
|||
|
|
" <td>0</td>\n",
|
|||
|
|
" <td>SSD</td>\n",
|
|||
|
|
" <td>No</td>\n",
|
|||
|
|
" <td>AMD</td>\n",
|
|||
|
|
" <td>Radeon Pro 455</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>4</th>\n",
|
|||
|
|
" <td>Apple</td>\n",
|
|||
|
|
" <td>MacBook Pro</td>\n",
|
|||
|
|
" <td>Ultrabook</td>\n",
|
|||
|
|
" <td>13.3</td>\n",
|
|||
|
|
" <td>8</td>\n",
|
|||
|
|
" <td>macOS</td>\n",
|
|||
|
|
" <td>1.37</td>\n",
|
|||
|
|
" <td>1803.60</td>\n",
|
|||
|
|
" <td>Standard</td>\n",
|
|||
|
|
" <td>2560</td>\n",
|
|||
|
|
" <td>...</td>\n",
|
|||
|
|
" <td>Yes</td>\n",
|
|||
|
|
" <td>Intel</td>\n",
|
|||
|
|
" <td>3.1</td>\n",
|
|||
|
|
" <td>Core i5</td>\n",
|
|||
|
|
" <td>256</td>\n",
|
|||
|
|
" <td>0</td>\n",
|
|||
|
|
" <td>SSD</td>\n",
|
|||
|
|
" <td>No</td>\n",
|
|||
|
|
" <td>Intel</td>\n",
|
|||
|
|
" <td>Iris Plus Graphics 650</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" </tbody>\n",
|
|||
|
|
"</table>\n",
|
|||
|
|
"<p>5 rows × 23 columns</p>\n",
|
|||
|
|
"</div>"
|
|||
|
|
],
|
|||
|
|
"text/plain": [
|
|||
|
|
" Company Product TypeName Inches Ram OS Weight Price_euros \\\n",
|
|||
|
|
"0 Apple MacBook Pro Ultrabook 13.3 8 macOS 1.37 1339.69 \n",
|
|||
|
|
"1 Apple Macbook Air Ultrabook 13.3 8 macOS 1.34 898.94 \n",
|
|||
|
|
"2 HP 250 G6 Notebook 15.6 8 No OS 1.86 575.00 \n",
|
|||
|
|
"3 Apple MacBook Pro Ultrabook 15.4 16 macOS 1.83 2537.45 \n",
|
|||
|
|
"4 Apple MacBook Pro Ultrabook 13.3 8 macOS 1.37 1803.60 \n",
|
|||
|
|
"\n",
|
|||
|
|
" Screen ScreenW ... RetinaDisplay CPU_company CPU_freq CPU_model \\\n",
|
|||
|
|
"0 Standard 2560 ... Yes Intel 2.3 Core i5 \n",
|
|||
|
|
"1 Standard 1440 ... No Intel 1.8 Core i5 \n",
|
|||
|
|
"2 Full HD 1920 ... No Intel 2.5 Core i5 7200U \n",
|
|||
|
|
"3 Standard 2880 ... Yes Intel 2.7 Core i7 \n",
|
|||
|
|
"4 Standard 2560 ... Yes Intel 3.1 Core i5 \n",
|
|||
|
|
"\n",
|
|||
|
|
" PrimaryStorage SecondaryStorage PrimaryStorageType SecondaryStorageType \\\n",
|
|||
|
|
"0 128 0 SSD No \n",
|
|||
|
|
"1 128 0 Flash Storage No \n",
|
|||
|
|
"2 256 0 SSD No \n",
|
|||
|
|
"3 512 0 SSD No \n",
|
|||
|
|
"4 256 0 SSD No \n",
|
|||
|
|
"\n",
|
|||
|
|
" GPU_company GPU_model \n",
|
|||
|
|
"0 Intel Iris Plus Graphics 640 \n",
|
|||
|
|
"1 Intel HD Graphics 6000 \n",
|
|||
|
|
"2 Intel HD Graphics 620 \n",
|
|||
|
|
"3 AMD Radeon Pro 455 \n",
|
|||
|
|
"4 Intel Iris Plus Graphics 650 \n",
|
|||
|
|
"\n",
|
|||
|
|
"[5 rows x 23 columns]"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"execution_count": 6,
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "execute_result"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Для наглядности\n",
|
|||
|
|
"df.head()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 7,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"text/html": [
|
|||
|
|
"<div>\n",
|
|||
|
|
"<style scoped>\n",
|
|||
|
|
" .dataframe tbody tr th:only-of-type {\n",
|
|||
|
|
" vertical-align: middle;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"\n",
|
|||
|
|
" .dataframe tbody tr th {\n",
|
|||
|
|
" vertical-align: top;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"\n",
|
|||
|
|
" .dataframe thead th {\n",
|
|||
|
|
" text-align: right;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"</style>\n",
|
|||
|
|
"<table border=\"1\" class=\"dataframe\">\n",
|
|||
|
|
" <thead>\n",
|
|||
|
|
" <tr style=\"text-align: right;\">\n",
|
|||
|
|
" <th></th>\n",
|
|||
|
|
" <th>Inches</th>\n",
|
|||
|
|
" <th>Ram</th>\n",
|
|||
|
|
" <th>Weight</th>\n",
|
|||
|
|
" <th>Price_euros</th>\n",
|
|||
|
|
" <th>ScreenW</th>\n",
|
|||
|
|
" <th>ScreenH</th>\n",
|
|||
|
|
" <th>CPU_freq</th>\n",
|
|||
|
|
" <th>PrimaryStorage</th>\n",
|
|||
|
|
" <th>SecondaryStorage</th>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" </thead>\n",
|
|||
|
|
" <tbody>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>count</th>\n",
|
|||
|
|
" <td>1275.000000</td>\n",
|
|||
|
|
" <td>1275.000000</td>\n",
|
|||
|
|
" <td>1275.000000</td>\n",
|
|||
|
|
" <td>1275.000000</td>\n",
|
|||
|
|
" <td>1275.000000</td>\n",
|
|||
|
|
" <td>1275.000000</td>\n",
|
|||
|
|
" <td>1275.000000</td>\n",
|
|||
|
|
" <td>1275.000000</td>\n",
|
|||
|
|
" <td>1275.000000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>mean</th>\n",
|
|||
|
|
" <td>15.022902</td>\n",
|
|||
|
|
" <td>8.440784</td>\n",
|
|||
|
|
" <td>2.040525</td>\n",
|
|||
|
|
" <td>1134.969059</td>\n",
|
|||
|
|
" <td>1900.043922</td>\n",
|
|||
|
|
" <td>1073.904314</td>\n",
|
|||
|
|
" <td>2.302980</td>\n",
|
|||
|
|
" <td>444.517647</td>\n",
|
|||
|
|
" <td>176.069020</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>std</th>\n",
|
|||
|
|
" <td>1.429470</td>\n",
|
|||
|
|
" <td>5.097809</td>\n",
|
|||
|
|
" <td>0.669196</td>\n",
|
|||
|
|
" <td>700.752504</td>\n",
|
|||
|
|
" <td>493.346186</td>\n",
|
|||
|
|
" <td>283.883940</td>\n",
|
|||
|
|
" <td>0.503846</td>\n",
|
|||
|
|
" <td>365.537726</td>\n",
|
|||
|
|
" <td>415.960655</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>min</th>\n",
|
|||
|
|
" <td>10.100000</td>\n",
|
|||
|
|
" <td>2.000000</td>\n",
|
|||
|
|
" <td>0.690000</td>\n",
|
|||
|
|
" <td>174.000000</td>\n",
|
|||
|
|
" <td>1366.000000</td>\n",
|
|||
|
|
" <td>768.000000</td>\n",
|
|||
|
|
" <td>0.900000</td>\n",
|
|||
|
|
" <td>8.000000</td>\n",
|
|||
|
|
" <td>0.000000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>25%</th>\n",
|
|||
|
|
" <td>14.000000</td>\n",
|
|||
|
|
" <td>4.000000</td>\n",
|
|||
|
|
" <td>1.500000</td>\n",
|
|||
|
|
" <td>609.000000</td>\n",
|
|||
|
|
" <td>1920.000000</td>\n",
|
|||
|
|
" <td>1080.000000</td>\n",
|
|||
|
|
" <td>2.000000</td>\n",
|
|||
|
|
" <td>256.000000</td>\n",
|
|||
|
|
" <td>0.000000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>50%</th>\n",
|
|||
|
|
" <td>15.600000</td>\n",
|
|||
|
|
" <td>8.000000</td>\n",
|
|||
|
|
" <td>2.040000</td>\n",
|
|||
|
|
" <td>989.000000</td>\n",
|
|||
|
|
" <td>1920.000000</td>\n",
|
|||
|
|
" <td>1080.000000</td>\n",
|
|||
|
|
" <td>2.500000</td>\n",
|
|||
|
|
" <td>256.000000</td>\n",
|
|||
|
|
" <td>0.000000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>75%</th>\n",
|
|||
|
|
" <td>15.600000</td>\n",
|
|||
|
|
" <td>8.000000</td>\n",
|
|||
|
|
" <td>2.310000</td>\n",
|
|||
|
|
" <td>1496.500000</td>\n",
|
|||
|
|
" <td>1920.000000</td>\n",
|
|||
|
|
" <td>1080.000000</td>\n",
|
|||
|
|
" <td>2.700000</td>\n",
|
|||
|
|
" <td>512.000000</td>\n",
|
|||
|
|
" <td>0.000000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>max</th>\n",
|
|||
|
|
" <td>18.400000</td>\n",
|
|||
|
|
" <td>64.000000</td>\n",
|
|||
|
|
" <td>4.700000</td>\n",
|
|||
|
|
" <td>6099.000000</td>\n",
|
|||
|
|
" <td>3840.000000</td>\n",
|
|||
|
|
" <td>2160.000000</td>\n",
|
|||
|
|
" <td>3.600000</td>\n",
|
|||
|
|
" <td>2048.000000</td>\n",
|
|||
|
|
" <td>2048.000000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" </tbody>\n",
|
|||
|
|
"</table>\n",
|
|||
|
|
"</div>"
|
|||
|
|
],
|
|||
|
|
"text/plain": [
|
|||
|
|
" Inches Ram Weight Price_euros ScreenW \\\n",
|
|||
|
|
"count 1275.000000 1275.000000 1275.000000 1275.000000 1275.000000 \n",
|
|||
|
|
"mean 15.022902 8.440784 2.040525 1134.969059 1900.043922 \n",
|
|||
|
|
"std 1.429470 5.097809 0.669196 700.752504 493.346186 \n",
|
|||
|
|
"min 10.100000 2.000000 0.690000 174.000000 1366.000000 \n",
|
|||
|
|
"25% 14.000000 4.000000 1.500000 609.000000 1920.000000 \n",
|
|||
|
|
"50% 15.600000 8.000000 2.040000 989.000000 1920.000000 \n",
|
|||
|
|
"75% 15.600000 8.000000 2.310000 1496.500000 1920.000000 \n",
|
|||
|
|
"max 18.400000 64.000000 4.700000 6099.000000 3840.000000 \n",
|
|||
|
|
"\n",
|
|||
|
|
" ScreenH CPU_freq PrimaryStorage SecondaryStorage \n",
|
|||
|
|
"count 1275.000000 1275.000000 1275.000000 1275.000000 \n",
|
|||
|
|
"mean 1073.904314 2.302980 444.517647 176.069020 \n",
|
|||
|
|
"std 283.883940 0.503846 365.537726 415.960655 \n",
|
|||
|
|
"min 768.000000 0.900000 8.000000 0.000000 \n",
|
|||
|
|
"25% 1080.000000 2.000000 256.000000 0.000000 \n",
|
|||
|
|
"50% 1080.000000 2.500000 256.000000 0.000000 \n",
|
|||
|
|
"75% 1080.000000 2.700000 512.000000 0.000000 \n",
|
|||
|
|
"max 2160.000000 3.600000 2048.000000 2048.000000 "
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"execution_count": 7,
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "execute_result"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Описание данных (основные статистические показатели)\n",
|
|||
|
|
"df.describe()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 8,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"text/plain": [
|
|||
|
|
"Company 0\n",
|
|||
|
|
"Product 0\n",
|
|||
|
|
"TypeName 0\n",
|
|||
|
|
"Inches 0\n",
|
|||
|
|
"Ram 0\n",
|
|||
|
|
"OS 0\n",
|
|||
|
|
"Weight 0\n",
|
|||
|
|
"Price_euros 0\n",
|
|||
|
|
"Screen 0\n",
|
|||
|
|
"ScreenW 0\n",
|
|||
|
|
"ScreenH 0\n",
|
|||
|
|
"Touchscreen 0\n",
|
|||
|
|
"IPSpanel 0\n",
|
|||
|
|
"RetinaDisplay 0\n",
|
|||
|
|
"CPU_company 0\n",
|
|||
|
|
"CPU_freq 0\n",
|
|||
|
|
"CPU_model 0\n",
|
|||
|
|
"PrimaryStorage 0\n",
|
|||
|
|
"SecondaryStorage 0\n",
|
|||
|
|
"PrimaryStorageType 0\n",
|
|||
|
|
"SecondaryStorageType 0\n",
|
|||
|
|
"GPU_company 0\n",
|
|||
|
|
"GPU_model 0\n",
|
|||
|
|
"dtype: int64"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"execution_count": 8,
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "execute_result"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Процент пропущенных значений признаков\n",
|
|||
|
|
"for i in df.columns:\n",
|
|||
|
|
" null_rate = df[i].isnull().sum() / len(df) * 100\n",
|
|||
|
|
" if null_rate > 0:\n",
|
|||
|
|
" print(f'{i} Процент пустых значений: %{null_rate:.2f}')\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Проверка на пропущенные данные\n",
|
|||
|
|
"df.isnull().sum()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 9,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"text/plain": [
|
|||
|
|
"Company False\n",
|
|||
|
|
"Product False\n",
|
|||
|
|
"TypeName False\n",
|
|||
|
|
"Inches False\n",
|
|||
|
|
"Ram False\n",
|
|||
|
|
"OS False\n",
|
|||
|
|
"Weight False\n",
|
|||
|
|
"Price_euros False\n",
|
|||
|
|
"Screen False\n",
|
|||
|
|
"ScreenW False\n",
|
|||
|
|
"ScreenH False\n",
|
|||
|
|
"Touchscreen False\n",
|
|||
|
|
"IPSpanel False\n",
|
|||
|
|
"RetinaDisplay False\n",
|
|||
|
|
"CPU_company False\n",
|
|||
|
|
"CPU_freq False\n",
|
|||
|
|
"CPU_model False\n",
|
|||
|
|
"PrimaryStorage False\n",
|
|||
|
|
"SecondaryStorage False\n",
|
|||
|
|
"PrimaryStorageType False\n",
|
|||
|
|
"SecondaryStorageType False\n",
|
|||
|
|
"GPU_company False\n",
|
|||
|
|
"GPU_model False\n",
|
|||
|
|
"dtype: bool"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"execution_count": 9,
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "execute_result"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"df.isnull().any()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"Пропущенных данных в колонках нет :)"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"## Бизнес-цель\n",
|
|||
|
|
"1. Оптимизация ценообразования \n",
|
|||
|
|
"Использование данных для построения модели прогнозирования цены ноутбуков на основе характеристик. Эффект: Повышение продаж за счет правильного позиционирования товаров в разных ценовых сегментах.\n",
|
|||
|
|
"2. Анализ рынка и планирование производства \n",
|
|||
|
|
"Определение наиболее популярных и прибыльных конфигураций ноутбуков на основе данных о характеристиках и продажах. Эффект: Снижение издержек на производство непопулярных моделей и повышение прибыльности.\n",
|
|||
|
|
"\n",
|
|||
|
|
"### Техническая цель\n",
|
|||
|
|
"1. Прогнозирование цены ноутбуков \n",
|
|||
|
|
"Создать модель машинного обучения, которая на вход получает характеристики ноутбука (ОЗУ, CPU, объём памяти, экран и т.д.) и предсказывает его цену. \n",
|
|||
|
|
"На вход поступают характеристики, а целевой признак - цена ноутбука. \n",
|
|||
|
|
"2. Классификация сегмента ноутбука \n",
|
|||
|
|
"Создать модель для предсказания ценового сегмента в зависимости от их характеристик. На вход поступают характеристики ноутбука, а целевой признак - определение ценового сегмента (бюджетный, средний, премиум)."
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 10,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1000x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1000x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Визуализация выбросов для столбца Price_euros\n",
|
|||
|
|
"plt.figure(figsize=(10, 6))\n",
|
|||
|
|
"sns.boxplot(x=df['Price_euros'])\n",
|
|||
|
|
"plt.title('Boxplot для цены ноутбуков (Price_euros)')\n",
|
|||
|
|
"plt.show()\n",
|
|||
|
|
"\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Визуализация распределения моделей процессоров\n",
|
|||
|
|
"plt.figure(figsize=(10, 6))\n",
|
|||
|
|
"df['CPU_model'].value_counts().plot(kind='bar')\n",
|
|||
|
|
"plt.title('Распределение процессоров ноутбуков')\n",
|
|||
|
|
"plt.xlabel('Модель процессора')\n",
|
|||
|
|
"plt.ylabel('Количество')\n",
|
|||
|
|
"plt.xticks(rotation=90)\n",
|
|||
|
|
"plt.show()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"### Проблема наборов \n",
|
|||
|
|
"\n",
|
|||
|
|
"**Зашумленность данных** — это наличие ненужной информации, ошибок или неверных значений, которые могут исказить анализ.\n",
|
|||
|
|
"Например, в поле OS могут быть пробелы или некорректные названия операционных систем (например, \"No OS\", \"Windows\", \"Windows 10\").\n",
|
|||
|
|
"\n",
|
|||
|
|
"**Смещение данных** может возникнуть, если данные отражают предпочтения определённого производителя или сегмента. Если в данных преобладает один бренд (например, Apple), это может сместить модель в сторону этого производителя.\n",
|
|||
|
|
"\n",
|
|||
|
|
"**Актуальность данных** - неактуальные данные могут влиять на точность прогнозирования. Ноутбуки устаревают быстро, поэтому информация о старых моделях может быть неактуальной для анализа современных трендов.\n",
|
|||
|
|
"\n",
|
|||
|
|
"**Выбросы** — это значения, которые значительно отличаются от большинства данных. Поле Price_euros может содержать аномально высокие или низкие цены, которые не соответствуют реальности.\n",
|
|||
|
|
"\n",
|
|||
|
|
"**Просачивание данных** - это когда информация из тестовой выборки \"просачивается\" в обучение, что даёт модели несправедливое преимущество.\n",
|
|||
|
|
"\n",
|
|||
|
|
"Решение: Очистить данные, убрать пробелы, проверить на корректность. Чётко разделить данные для обучения и тестирования, исключить целевые признаки из обучения. Уравнять количество данных для разных производителей или использовать методы балансировки.\n",
|
|||
|
|
"\n",
|
|||
|
|
"Для устранения смещения применим рэсемплинг данных, либо же балансировку."
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 12,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1000x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1000x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Балансировка данных. Визуализация до ресэмплинга\n",
|
|||
|
|
"plt.figure(figsize=(10, 6))\n",
|
|||
|
|
"df['Company'].value_counts().plot(kind='bar', color='blue')\n",
|
|||
|
|
"plt.title('Распределение компаний до ресэмплинга')\n",
|
|||
|
|
"plt.xlabel('Компания')\n",
|
|||
|
|
"plt.ylabel('Количество')\n",
|
|||
|
|
"plt.xticks(rotation=45)\n",
|
|||
|
|
"plt.show()\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Пример ресэмплинга\n",
|
|||
|
|
"brand_counts = df['Company'].value_counts()\n",
|
|||
|
|
"min_count = brand_counts.min()\n",
|
|||
|
|
"balanced_data = pd.concat([df[df['Company'] == brand].sample(min_count) for brand in brand_counts.index])\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Визуализация после ресэмплинга\n",
|
|||
|
|
"plt.figure(figsize=(10, 6))\n",
|
|||
|
|
"balanced_data['Company'].value_counts().plot(kind='bar', color='green')\n",
|
|||
|
|
"plt.title('Распределение компаний после ресэмплинга')\n",
|
|||
|
|
"plt.xlabel('Компания')\n",
|
|||
|
|
"plt.ylabel('Количество')\n",
|
|||
|
|
"plt.xticks(rotation=45)\n",
|
|||
|
|
"plt.show()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"Рэсемплинг данных позволил нам изменить размер выборки, чтобы достичь более сбалансированного распределения классов, чтобы повысить общую эффективность анализа данных. Устанили перекос (дисбаланс) данных и добились более равномерное распределение компаний (брендов) ноутбуков."
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 35,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"name": "stdout",
|
|||
|
|
"output_type": "stream",
|
|||
|
|
"text": [
|
|||
|
|
"Количество выявленных выбросов: 101\n",
|
|||
|
|
"Количество записей после удаления выбросов: 1174\n"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"import numpy as np\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Вычисление медианы для 'Price_euros'\n",
|
|||
|
|
"median_price = df['Price_euros'].median()\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Вычисление абсолютных отклонений от медианы\n",
|
|||
|
|
"df['absolute_deviation'] = np.abs(df['Price_euros'] - median_price)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Вычисление медианного абсолютного отклонения (MAD)\n",
|
|||
|
|
"mad = df['absolute_deviation'].median()\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Определение порога для выбросов. Вы можете настроить множитель в зависимости от ваших предпочтений\n",
|
|||
|
|
"threshold = 3 * mad\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Выявление выбросов\n",
|
|||
|
|
"outliers = df[df['absolute_deviation'] > threshold]\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Удаление выбросов из набора данных\n",
|
|||
|
|
"datalaptop_cleaned = df[df['absolute_deviation'] <= threshold].drop(columns=['absolute_deviation'])\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Вывод количества найденных выбросов и оставшихся записей\n",
|
|||
|
|
"print(f\"Количество выявленных выбросов: {len(outliers)}\")\n",
|
|||
|
|
"print(f\"Количество записей после удаления выбросов: {len(datalaptop_cleaned)}\")\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Визуализация результатов\n",
|
|||
|
|
"plt.figure(figsize=(12, 6))\n",
|
|||
|
|
"sns.boxplot(x=df['Price_euros'])\n",
|
|||
|
|
"plt.title('Распределение Price_euros с выбросами')\n",
|
|||
|
|
"plt.show()\n",
|
|||
|
|
"\n",
|
|||
|
|
"plt.figure(figsize=(12, 6))\n",
|
|||
|
|
"sns.boxplot(x=datalaptop_cleaned['Price_euros'])\n",
|
|||
|
|
"plt.title('Распределение Price_euros после удаления выбросов')\n",
|
|||
|
|
"plt.show()\n"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"Медиальный способ борьбы с выбросами позволяет сохранить структуру даныых, создать устойчивость к этим выбросать, чтобы избежать смещения данных, если использовали бы среднее. Если имеются товары с высоко аномальной ценой, то замена на медиану позволит сократить влияние на данные. Удаление выбросов вдобавок и решает проблему зашумлённости данных. Выбросы - это одна из форм шума и удаление их позволяеи сделать данные чище и менее зашумлённые."
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 47,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"name": "stdout",
|
|||
|
|
"output_type": "stream",
|
|||
|
|
"text": [
|
|||
|
|
"Размер обучающей выборки: 765\n",
|
|||
|
|
"Размер валидационной выборки: 255\n",
|
|||
|
|
"Размер тестовой выборки: 255\n",
|
|||
|
|
"Company float64\n",
|
|||
|
|
"Product float64\n",
|
|||
|
|
"TypeName float64\n",
|
|||
|
|
"Inches float64\n",
|
|||
|
|
"Ram int64\n",
|
|||
|
|
"OS float64\n",
|
|||
|
|
"Weight float64\n",
|
|||
|
|
"Screen float64\n",
|
|||
|
|
"ScreenW int64\n",
|
|||
|
|
"ScreenH int64\n",
|
|||
|
|
"Touchscreen float64\n",
|
|||
|
|
"IPSpanel float64\n",
|
|||
|
|
"RetinaDisplay float64\n",
|
|||
|
|
"CPU_company float64\n",
|
|||
|
|
"CPU_freq float64\n",
|
|||
|
|
"CPU_model float64\n",
|
|||
|
|
"PrimaryStorage int64\n",
|
|||
|
|
"SecondaryStorage int64\n",
|
|||
|
|
"PrimaryStorageType float64\n",
|
|||
|
|
"SecondaryStorageType float64\n",
|
|||
|
|
"GPU_company float64\n",
|
|||
|
|
"GPU_model float64\n",
|
|||
|
|
"dtype: object\n"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "iVBORw0KGgoAAAANSUhEUgAAA1IAAAIjCAYAAAAJLyrXAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAAB1zUlEQVR4nO3dd3gU5f7+8XvTQyoB0iAkdBBCC0V6FyKgCMqhKKBIR49ixQaiR+zlKFKOCiggSEdUkBpEEWmhdwOhJdT0nszvj/zYL0tCmRhIAu/Xde0F+0z77E5md++ZZ2YshmEYAgAAAADcNLuiLgAAAAAAShqCFAAAAACYRJACAAAAAJMIUgAAAABgEkEKAAAAAEwiSAEAAACASQQpAAAAADCJIAUAAAAAJhGkAAAAAMAkghQAAABuibZt26pt27bW58eOHZPFYtGMGTOKrCagsBCkgEI0Y8YMWSwW68PFxUXVq1fX6NGjFRsbW9TlAQAAoJA4FHUBwJ1owoQJqlSpktLS0rRx40ZNnjxZP//8s/bs2aNSpUoVdXkAABSJ4OBgpaamytHRsahLAf4xghRwC4SHh6tRo0aSpCeffFJlypTRxx9/rKVLl6pv375FXB0AoKRITk6Wm5tbUZdRaC731gDuBHTtA26D9u3bS5KioqIkSRcvXtTzzz+v0NBQubu7y9PTU+Hh4dq5c2eeadPS0jR+/HhVr15dLi4uCggIUM+ePXX06FFJ/9ff/FqPK/umr1+/XhaLRfPmzdMrr7wif39/ubm56YEHHtCJEyfyLHvz5s3q0qWLvLy8VKpUKbVp00a///57vq+xbdu2+S5//PjxecadNWuWwsLC5OrqKh8fH/Xp0yff5V/vtV0pJydHn376qWrXri0XFxf5+flp2LBhunTpks14ISEh6tatW57ljB49Os8886v9gw8+yPOeSlJ6errGjRunqlWrytnZWUFBQXrxxReVnp6e73t1pbZt26pOnTp52j/88ENZLBYdO3bMpj0uLk7PPPOMgoKC5OzsrKpVq+q9995TTk5OnnmMHz8+3/du0KBBNuOdOnVKTzzxhPz8/OTs7KzatWvrm2++sRnn8t/O5Yezs7OqV6+uiRMnyjAMm3F37Nih8PBweXp6yt3dXR06dNCff/5pM87V3WCvfJw8edI63tq1a9WqVSu5ubnJ29tbDz74oPbv33/D9/Xqeq/3HlyuZcOGDRo2bJjKlCkjT09PDRgwIN+/oavfv/nz58tisSgkJMTadvDgQbVv317+/v7Wv4nhw4fr4sWLeWpcsGBBnvrd3d1tlnOznxmX57l+/Xpr2+nTpxUSEqJGjRopKSlJkpSRkaE33nhDYWFh8vLykpubm1q1aqV169bd8L29/D5c7/29UlZWlt566y1VqVJFzs7OCgkJ0SuvvJLv9vHLL7+oTZs28vDwkKenpxo3bqw5c+bYjFPYnwvXcqO/vQULFshisSgiIiLPtFOnTpXFYtGePXusbQcOHNDDDz8sHx8fubi4qFGjRlq2bJnNdJf/FiMiIjRy5Ej5+vqqQoUKkqTExEQ988wzCgkJkbOzs3x9fdWpUydt377dOv1vv/2mRx55RBUrVrT+3T377LNKTU21Wc6gQYPk7u6u6OhodevWTe7u7ipfvrwmTZokSdq9e7fat28vNzc3BQcH51kHZraZq+V3jtTlek6dOqUePXrI3d1d5cqV0/PPP6/s7Gyb6S9cuKDHHntMnp6e8vb21sCBA7Vz507Ou0KR4IgUcBtcDj1lypSRJP39999asmSJHnnkEVWqVEmxsbGaOnWq2rRpo3379ikwMFCSlJ2drW7dumnNmjXq06eP/v3vfysxMVGrVq3Snj17VKVKFesy+vbtq/vvv99muWPHjs23nv/85z+yWCx66aWXdPbsWX366afq2LGjIiMj5erqKin3R0R4eLjCwsI0btw42dnZafr06Wrfvr1+++03NWnSJM98K1SooIkTJ0qSkpKSNGLEiHyX/frrr6t379568sknde7cOX3++edq3bq1duzYIW9v7zzTDB06VK1atZIkLVq0SIsXL7YZPmzYMM2YMUOPP/64nn76aUVFRemLL77Qjh079PvvvxdKF5K4uDjra7tSTk6OHnjgAW3cuFFDhw5VrVq1tHv3bn3yySc6dOiQlixZ8o+XfVlKSoratGmjU6dOadiwYapYsaL++OMPjR07VmfOnNGnn36a73Tfffed9f/PPvuszbDY2Fjde++9slgsGj16tMqVK6dffvlFgwcPVkJCgp555hmb8V955RXVqlVLqamp1kDu6+urwYMHS5L27t2rVq1aydPTUy+++KIcHR01depUtW3bVhEREWratKnN/C53g72Sj4+PJGn16tUKDw9X5cqVNX78eKWmpurzzz9XixYttH37dpvgci1PP/20GjdubNP25JNP5jvu6NGj5e3trfHjx+vgwYOaPHmyjh8/bg0n+cnKytKrr76apz05OVkVKlRQ9+7d5enpqT179mjSpEk6deqUfvzxxxvWfbWb/cy4Wnx8vMLDw+Xo6Kiff/5Z7u7ukqSEhAR99dVX6tu3r4YMGaLExER9/fXX6ty5s/766y/Vr1//hjXVr19fzz33nE3bt99+q1WrVtm0Pfnkk5o5c6YefvhhPffcc9q8ebMmTpyo/fv322zLM2bM0BNPPKHatWtr7Nix8vb21o4dO7RixQr169cvz/Jv5efCzfztde3aVe7u7vrhhx/Upk0bm+nnzZun2rVrW3eS7N27Vy1atFD58uX18ssvy83NTT/88IN69OihhQsX6qGHHrKZfuTIkSpXrpzeeOMNJScnS5KGDx+uBQsWaPTo0brnnnt04cIFbdy4Ufv371fDhg0l5Yb6lJQUjRgxQmXKlNFff/2lzz//XCdPntT8+fNtlpGdna3w8HC1bt1a77//vmbPnq3Ro0fLzc1Nr776qvr376+ePXtqypQpGjBggJo1a5ZnWy3INnMt2dnZ6ty5s5o2baoPP/xQq1ev1kcffaQqVapYv0tycnLUvXt3/fXXXxoxYoRq1qyppUuXauDAgaaWBRQaA0ChmT59uiHJWL16tXHu3DnjxIkTxty5c40yZcoYrq6uxsmTJw3DMIy0tDQjOzvbZtqoqCjD2dnZmDBhgrXtm2++MSQZH3/8cZ5l5eTkWKeTZHzwwQd5xqldu7bRpk0b6/N169YZkozy5csbCQkJ1vYffvjBkGR89tln1nlXq1bN6Ny5s3U5hmEYKSkpRqVKlYxOnTrlWVbz5s2NOnXqWJ+fO3fOkGSMGzfO2nbs2DHD3t7e+M9//mMz7e7duw0HB4c87YcPHzYkGTNnzrS2jRs3zrjyo+u3334zJBmzZ8+2mXbFihV52oODg42uXbvmqX3UqFHG1R+HV9f+4osvGr6+vkZYWJjNe/rdd98ZdnZ2xm+//WYz/ZQpUwxJxu+//55neVdq06aNUbt27TztH3zwgSHJiIqKsra99dZbhpubm3Ho0CGbcV9++WXD3t7eiI6Otml/9dVXDYvFYtMWHBxsDBw40Pp88ODBRkBAgHH+/Hmb8fr06WN4eXkZKSkphmH839/OunXrrOOkpaUZdnZ2xsiRI61tPXr0MJycnIyjR49a206fPm14eHgYrVu3trZd3la2bNlyjXfGMOrXr2/4+voaFy5csLbt3LnTsLOzMwYMGHDN6a6sd/78+XmGubm52bwHl2sJCwszMjIyrO3vv/++IclYunSpte3q9+/LL780nJ2djXbt2hnBwcHXrWnkyJGGu7t7gWq82c+MK9dTWlqa0bZtW8PX19c4cuSIzbRZWVlGenq6TdulS5cMPz8/44knnrju6zCMm9+WIiMjDUnGk08+aTPe888/b0gy1q5daxiGYcTFxRkeHh5G06ZNjdTUVJtxr/wMMozC/1zIz83+7fXt29fw9fU1srKyrG1nzpwx7OzsbNZLhw4djNDQUCMtLc3mdTVv3tyoVq2ate3y32LLli1t5mkYhuHl5WWMGjXqunVf3l6vNHHiRMNisRjHjx+3tg0cONCQZLzzzjvWtkuXLhmurq6
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1000x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1000x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"name": "stdout",
|
|||
|
|
"output_type": "stream",
|
|||
|
|
"text": [
|
|||
|
|
"Количество записей в обучающей выборке до аугментации: 765\n",
|
|||
|
|
"Количество записей в обучающей выборке после oversampling: 772\n",
|
|||
|
|
"Количество записей в обучающей выборке после undersampling: 760\n"
|
|||
|
|
]
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"from imblearn.over_sampling import RandomOverSampler\n",
|
|||
|
|
"from imblearn.under_sampling import RandomUnderSampler\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Разделение признаков и целевой переменной\n",
|
|||
|
|
"X = df.drop(columns=['Price_euros']) # Признаки\n",
|
|||
|
|
"y = df['Price_euros'] # Целевая переменная (для регрессии)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Разбиение данных на обучающую, контрольную и тестовую выборки\n",
|
|||
|
|
"X_train, X_temp, y_train, y_temp = train_test_split(X, y, test_size=0.4, random_state=42)\n",
|
|||
|
|
"X_val, X_test, y_val, y_test = train_test_split(X_temp, y_temp, test_size=0.5, random_state=42)\n",
|
|||
|
|
"\n",
|
|||
|
|
"print(f\"Размер обучающей выборки: {len(X_train)}\")\n",
|
|||
|
|
"print(f\"Размер валидационной выборки: {len(X_val)}\")\n",
|
|||
|
|
"print(f\"Размер тестовой выборки: {len(X_test)}\")\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Все типы к числовым приводим\n",
|
|||
|
|
"X_train = X_train.apply(pd.to_numeric, errors='coerce')\n",
|
|||
|
|
"print(X_train.dtypes)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Квантилизация целевой переменной для создания категорий\n",
|
|||
|
|
"y_train_cat = pd.qcut(y_train, q=4, labels=False)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Объединяем исходные данные и \"шумные\" данные для увеличения обучающей выборки\n",
|
|||
|
|
"X_train_combined = np.vstack([X_train, X_train])\n",
|
|||
|
|
"y_train_combined = np.hstack([y_train, y_train]) # Убедитесь, что y_train повторяется для новых данных\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Применение oversampling и undersampling\n",
|
|||
|
|
"ros = RandomOverSampler(random_state=42)\n",
|
|||
|
|
"X_train_ros, y_train_ros = ros.fit_resample(X_train, y_train_cat)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Проверка распределения после oversampling\n",
|
|||
|
|
"plt.figure(figsize=(10, 6))\n",
|
|||
|
|
"sns.histplot(y_train_ros, bins=30, kde=True, color='green')\n",
|
|||
|
|
"plt.title('Распределение целевого признака после oversampling')\n",
|
|||
|
|
"plt.xlabel('Price_euros')\n",
|
|||
|
|
"plt.ylabel('Частота')\n",
|
|||
|
|
"plt.show()\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Применение RandomUnderSampler\n",
|
|||
|
|
"rus = RandomUnderSampler(random_state=42)\n",
|
|||
|
|
"X_train_rus, y_train_rus = rus.fit_resample(X_train, y_train_cat)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Проверка распределения после undersampling\n",
|
|||
|
|
"plt.figure(figsize=(10, 6))\n",
|
|||
|
|
"sns.histplot(y_train_rus, bins=30, kde=True, color='red')\n",
|
|||
|
|
"plt.title('Распределение целевого признака после undersampling')\n",
|
|||
|
|
"plt.xlabel('Price_euros')\n",
|
|||
|
|
"plt.ylabel('Частота')\n",
|
|||
|
|
"plt.show()\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Вывод количества данных до и после аугментации\n",
|
|||
|
|
"print(f\"Количество записей в обучающей выборке до аугментации: {len(y_train)}\")\n",
|
|||
|
|
"print(f\"Количество записей в обучающей выборке после oversampling: {len(y_train_ros)}\")\n",
|
|||
|
|
"print(f\"Количество записей в обучающей выборке после undersampling: {len(y_train_rus)}\")"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"Набор данных, судя по числам, был относительно сбалансирован, и применение oversampling/undersampling не сильно изменил размер выборок. Использование RandomOverSampler и RandomUnderSampler для регрессионных задач не является оптимальным решением, поскольку эти методы предназначены для классификации."
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"# Третий датафрейм: Зарплата сотрудников по обработке данных (DataScience)\n",
|
|||
|
|
"Выгрузили третий датасет!!! :)"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 48,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"name": "stdout",
|
|||
|
|
"output_type": "stream",
|
|||
|
|
"text": [
|
|||
|
|
"Index(['work_year', 'experience_level', 'employment_type', 'job_title',\n",
|
|||
|
|
" 'salary', 'salary_currency', 'salary_in_usd', 'employee_residence',\n",
|
|||
|
|
" 'remote_ratio', 'company_location', 'company_size'],\n",
|
|||
|
|
" dtype='object')\n",
|
|||
|
|
"\n",
|
|||
|
|
"<class 'pandas.core.frame.DataFrame'>\n",
|
|||
|
|
"RangeIndex: 3755 entries, 0 to 3754\n",
|
|||
|
|
"Data columns (total 11 columns):\n",
|
|||
|
|
" # Column Non-Null Count Dtype \n",
|
|||
|
|
"--- ------ -------------- ----- \n",
|
|||
|
|
" 0 work_year 3755 non-null int64 \n",
|
|||
|
|
" 1 experience_level 3755 non-null object\n",
|
|||
|
|
" 2 employment_type 3755 non-null object\n",
|
|||
|
|
" 3 job_title 3755 non-null object\n",
|
|||
|
|
" 4 salary 3755 non-null int64 \n",
|
|||
|
|
" 5 salary_currency 3755 non-null object\n",
|
|||
|
|
" 6 salary_in_usd 3755 non-null int64 \n",
|
|||
|
|
" 7 employee_residence 3755 non-null object\n",
|
|||
|
|
" 8 remote_ratio 3755 non-null int64 \n",
|
|||
|
|
" 9 company_location 3755 non-null object\n",
|
|||
|
|
" 10 company_size 3755 non-null object\n",
|
|||
|
|
"dtypes: int64(4), object(7)\n",
|
|||
|
|
"memory usage: 322.8+ KB\n"
|
|||
|
|
]
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Выгружаем третий датафрейм\n",
|
|||
|
|
"df = pd.read_csv(\".\\\\static\\\\csv\\\\ds_salaries.csv\")\n",
|
|||
|
|
"\n",
|
|||
|
|
"print(df.columns)\n",
|
|||
|
|
"\n",
|
|||
|
|
"print()\n",
|
|||
|
|
"\n",
|
|||
|
|
"df.info()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"## Описание набора \n",
|
|||
|
|
"Kaggle даёт следующие пояснения о датасете: \n",
|
|||
|
|
"Целью данного исследования является изучение факторов, влияющих на заработную плату специалистов по обработке данных. Для этого был использован набор данных, содержащий различные релевантные переменные. В этом отчёте описывается исследовательский анализ, проведённый для понимания взаимосвязи между этими факторами и заработной платой специалистов по обработке данных. Наука о данных — быстро развивающаяся область, и специалисты по обработке данных играют важнейшую роль в анализе и интерпретации больших объёмов данных. Поскольку эта профессия становится всё более востребованной, важно понимать факторы, которые могут влиять на зарплату специалистов по обработке данных. Этот анализ посвящён изучению этих факторов и их влиянию на зарплату."
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 49,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"text/html": [
|
|||
|
|
"<div>\n",
|
|||
|
|
"<style scoped>\n",
|
|||
|
|
" .dataframe tbody tr th:only-of-type {\n",
|
|||
|
|
" vertical-align: middle;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"\n",
|
|||
|
|
" .dataframe tbody tr th {\n",
|
|||
|
|
" vertical-align: top;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"\n",
|
|||
|
|
" .dataframe thead th {\n",
|
|||
|
|
" text-align: right;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"</style>\n",
|
|||
|
|
"<table border=\"1\" class=\"dataframe\">\n",
|
|||
|
|
" <thead>\n",
|
|||
|
|
" <tr style=\"text-align: right;\">\n",
|
|||
|
|
" <th></th>\n",
|
|||
|
|
" <th>work_year</th>\n",
|
|||
|
|
" <th>experience_level</th>\n",
|
|||
|
|
" <th>employment_type</th>\n",
|
|||
|
|
" <th>job_title</th>\n",
|
|||
|
|
" <th>salary</th>\n",
|
|||
|
|
" <th>salary_currency</th>\n",
|
|||
|
|
" <th>salary_in_usd</th>\n",
|
|||
|
|
" <th>employee_residence</th>\n",
|
|||
|
|
" <th>remote_ratio</th>\n",
|
|||
|
|
" <th>company_location</th>\n",
|
|||
|
|
" <th>company_size</th>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" </thead>\n",
|
|||
|
|
" <tbody>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>0</th>\n",
|
|||
|
|
" <td>2023</td>\n",
|
|||
|
|
" <td>SE</td>\n",
|
|||
|
|
" <td>FT</td>\n",
|
|||
|
|
" <td>Principal Data Scientist</td>\n",
|
|||
|
|
" <td>80000</td>\n",
|
|||
|
|
" <td>EUR</td>\n",
|
|||
|
|
" <td>85847</td>\n",
|
|||
|
|
" <td>ES</td>\n",
|
|||
|
|
" <td>100</td>\n",
|
|||
|
|
" <td>ES</td>\n",
|
|||
|
|
" <td>L</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>1</th>\n",
|
|||
|
|
" <td>2023</td>\n",
|
|||
|
|
" <td>MI</td>\n",
|
|||
|
|
" <td>CT</td>\n",
|
|||
|
|
" <td>ML Engineer</td>\n",
|
|||
|
|
" <td>30000</td>\n",
|
|||
|
|
" <td>USD</td>\n",
|
|||
|
|
" <td>30000</td>\n",
|
|||
|
|
" <td>US</td>\n",
|
|||
|
|
" <td>100</td>\n",
|
|||
|
|
" <td>US</td>\n",
|
|||
|
|
" <td>S</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>2</th>\n",
|
|||
|
|
" <td>2023</td>\n",
|
|||
|
|
" <td>MI</td>\n",
|
|||
|
|
" <td>CT</td>\n",
|
|||
|
|
" <td>ML Engineer</td>\n",
|
|||
|
|
" <td>25500</td>\n",
|
|||
|
|
" <td>USD</td>\n",
|
|||
|
|
" <td>25500</td>\n",
|
|||
|
|
" <td>US</td>\n",
|
|||
|
|
" <td>100</td>\n",
|
|||
|
|
" <td>US</td>\n",
|
|||
|
|
" <td>S</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>3</th>\n",
|
|||
|
|
" <td>2023</td>\n",
|
|||
|
|
" <td>SE</td>\n",
|
|||
|
|
" <td>FT</td>\n",
|
|||
|
|
" <td>Data Scientist</td>\n",
|
|||
|
|
" <td>175000</td>\n",
|
|||
|
|
" <td>USD</td>\n",
|
|||
|
|
" <td>175000</td>\n",
|
|||
|
|
" <td>CA</td>\n",
|
|||
|
|
" <td>100</td>\n",
|
|||
|
|
" <td>CA</td>\n",
|
|||
|
|
" <td>M</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>4</th>\n",
|
|||
|
|
" <td>2023</td>\n",
|
|||
|
|
" <td>SE</td>\n",
|
|||
|
|
" <td>FT</td>\n",
|
|||
|
|
" <td>Data Scientist</td>\n",
|
|||
|
|
" <td>120000</td>\n",
|
|||
|
|
" <td>USD</td>\n",
|
|||
|
|
" <td>120000</td>\n",
|
|||
|
|
" <td>CA</td>\n",
|
|||
|
|
" <td>100</td>\n",
|
|||
|
|
" <td>CA</td>\n",
|
|||
|
|
" <td>M</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" </tbody>\n",
|
|||
|
|
"</table>\n",
|
|||
|
|
"</div>"
|
|||
|
|
],
|
|||
|
|
"text/plain": [
|
|||
|
|
" work_year experience_level employment_type job_title \\\n",
|
|||
|
|
"0 2023 SE FT Principal Data Scientist \n",
|
|||
|
|
"1 2023 MI CT ML Engineer \n",
|
|||
|
|
"2 2023 MI CT ML Engineer \n",
|
|||
|
|
"3 2023 SE FT Data Scientist \n",
|
|||
|
|
"4 2023 SE FT Data Scientist \n",
|
|||
|
|
"\n",
|
|||
|
|
" salary salary_currency salary_in_usd employee_residence remote_ratio \\\n",
|
|||
|
|
"0 80000 EUR 85847 ES 100 \n",
|
|||
|
|
"1 30000 USD 30000 US 100 \n",
|
|||
|
|
"2 25500 USD 25500 US 100 \n",
|
|||
|
|
"3 175000 USD 175000 CA 100 \n",
|
|||
|
|
"4 120000 USD 120000 CA 100 \n",
|
|||
|
|
"\n",
|
|||
|
|
" company_location company_size \n",
|
|||
|
|
"0 ES L \n",
|
|||
|
|
"1 US S \n",
|
|||
|
|
"2 US S \n",
|
|||
|
|
"3 CA M \n",
|
|||
|
|
"4 CA M "
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"execution_count": 49,
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "execute_result"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Для наглядности\n",
|
|||
|
|
"df.head()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 50,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"text/html": [
|
|||
|
|
"<div>\n",
|
|||
|
|
"<style scoped>\n",
|
|||
|
|
" .dataframe tbody tr th:only-of-type {\n",
|
|||
|
|
" vertical-align: middle;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"\n",
|
|||
|
|
" .dataframe tbody tr th {\n",
|
|||
|
|
" vertical-align: top;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"\n",
|
|||
|
|
" .dataframe thead th {\n",
|
|||
|
|
" text-align: right;\n",
|
|||
|
|
" }\n",
|
|||
|
|
"</style>\n",
|
|||
|
|
"<table border=\"1\" class=\"dataframe\">\n",
|
|||
|
|
" <thead>\n",
|
|||
|
|
" <tr style=\"text-align: right;\">\n",
|
|||
|
|
" <th></th>\n",
|
|||
|
|
" <th>work_year</th>\n",
|
|||
|
|
" <th>salary</th>\n",
|
|||
|
|
" <th>salary_in_usd</th>\n",
|
|||
|
|
" <th>remote_ratio</th>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" </thead>\n",
|
|||
|
|
" <tbody>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>count</th>\n",
|
|||
|
|
" <td>3755.000000</td>\n",
|
|||
|
|
" <td>3.755000e+03</td>\n",
|
|||
|
|
" <td>3755.000000</td>\n",
|
|||
|
|
" <td>3755.000000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>mean</th>\n",
|
|||
|
|
" <td>2022.373635</td>\n",
|
|||
|
|
" <td>1.906956e+05</td>\n",
|
|||
|
|
" <td>137570.389880</td>\n",
|
|||
|
|
" <td>46.271638</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>std</th>\n",
|
|||
|
|
" <td>0.691448</td>\n",
|
|||
|
|
" <td>6.716765e+05</td>\n",
|
|||
|
|
" <td>63055.625278</td>\n",
|
|||
|
|
" <td>48.589050</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>min</th>\n",
|
|||
|
|
" <td>2020.000000</td>\n",
|
|||
|
|
" <td>6.000000e+03</td>\n",
|
|||
|
|
" <td>5132.000000</td>\n",
|
|||
|
|
" <td>0.000000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>25%</th>\n",
|
|||
|
|
" <td>2022.000000</td>\n",
|
|||
|
|
" <td>1.000000e+05</td>\n",
|
|||
|
|
" <td>95000.000000</td>\n",
|
|||
|
|
" <td>0.000000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>50%</th>\n",
|
|||
|
|
" <td>2022.000000</td>\n",
|
|||
|
|
" <td>1.380000e+05</td>\n",
|
|||
|
|
" <td>135000.000000</td>\n",
|
|||
|
|
" <td>0.000000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>75%</th>\n",
|
|||
|
|
" <td>2023.000000</td>\n",
|
|||
|
|
" <td>1.800000e+05</td>\n",
|
|||
|
|
" <td>175000.000000</td>\n",
|
|||
|
|
" <td>100.000000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" <tr>\n",
|
|||
|
|
" <th>max</th>\n",
|
|||
|
|
" <td>2023.000000</td>\n",
|
|||
|
|
" <td>3.040000e+07</td>\n",
|
|||
|
|
" <td>450000.000000</td>\n",
|
|||
|
|
" <td>100.000000</td>\n",
|
|||
|
|
" </tr>\n",
|
|||
|
|
" </tbody>\n",
|
|||
|
|
"</table>\n",
|
|||
|
|
"</div>"
|
|||
|
|
],
|
|||
|
|
"text/plain": [
|
|||
|
|
" work_year salary salary_in_usd remote_ratio\n",
|
|||
|
|
"count 3755.000000 3.755000e+03 3755.000000 3755.000000\n",
|
|||
|
|
"mean 2022.373635 1.906956e+05 137570.389880 46.271638\n",
|
|||
|
|
"std 0.691448 6.716765e+05 63055.625278 48.589050\n",
|
|||
|
|
"min 2020.000000 6.000000e+03 5132.000000 0.000000\n",
|
|||
|
|
"25% 2022.000000 1.000000e+05 95000.000000 0.000000\n",
|
|||
|
|
"50% 2022.000000 1.380000e+05 135000.000000 0.000000\n",
|
|||
|
|
"75% 2023.000000 1.800000e+05 175000.000000 100.000000\n",
|
|||
|
|
"max 2023.000000 3.040000e+07 450000.000000 100.000000"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"execution_count": 50,
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "execute_result"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Описание представленного датафрейма\n",
|
|||
|
|
"df.describe()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 51,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"text/plain": [
|
|||
|
|
"work_year 0\n",
|
|||
|
|
"experience_level 0\n",
|
|||
|
|
"employment_type 0\n",
|
|||
|
|
"job_title 0\n",
|
|||
|
|
"salary 0\n",
|
|||
|
|
"salary_currency 0\n",
|
|||
|
|
"salary_in_usd 0\n",
|
|||
|
|
"employee_residence 0\n",
|
|||
|
|
"remote_ratio 0\n",
|
|||
|
|
"company_location 0\n",
|
|||
|
|
"company_size 0\n",
|
|||
|
|
"dtype: int64"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"execution_count": 51,
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "execute_result"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Процент пропущенных значений признаков\n",
|
|||
|
|
"for i in df.columns:\n",
|
|||
|
|
" null_rate = df[i].isnull().sum() / len(df) * 100\n",
|
|||
|
|
" if null_rate > 0:\n",
|
|||
|
|
" print(f'{i} Процент пустых значений: %{null_rate:.2f}')\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Проверка на пропущенные данные\n",
|
|||
|
|
"df.isnull().sum()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 52,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"text/plain": [
|
|||
|
|
"work_year False\n",
|
|||
|
|
"experience_level False\n",
|
|||
|
|
"employment_type False\n",
|
|||
|
|
"job_title False\n",
|
|||
|
|
"salary False\n",
|
|||
|
|
"salary_currency False\n",
|
|||
|
|
"salary_in_usd False\n",
|
|||
|
|
"employee_residence False\n",
|
|||
|
|
"remote_ratio False\n",
|
|||
|
|
"company_location False\n",
|
|||
|
|
"company_size False\n",
|
|||
|
|
"dtype: bool"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"execution_count": 52,
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "execute_result"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"df.isnull().any()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"Хе-хе, какие весёлые датасеты :) Ни где данных пропущенных не нашлось, хе-хе))"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"# Анализ сведений о наборе данных\n",
|
|||
|
|
"\n",
|
|||
|
|
"**Набор данных:** Набор данных представляет собой информацию о сотрудниках DataScientist. Их зарплатах, занятости, стаже и по этим данным будет проводиться анализ. \n",
|
|||
|
|
"\n",
|
|||
|
|
"**Проблемная область:** Проблемной областью может являться анализ и управление заработной платой сотрудников с учётом различных факторов, которые позволяют компаниям улучшать свои кадровые стратегии и конкурировать на рынке труда.\n",
|
|||
|
|
"\n",
|
|||
|
|
"Проанализируем содержимое...\n",
|
|||
|
|
"\n",
|
|||
|
|
"*Объекты наблюдения:* Каждый сотрдник в области оьработки данных является объектом наблюдения. \n",
|
|||
|
|
"\n",
|
|||
|
|
"*Атрибуты объектов:* Основные атрибуты представляют собой зарплату работников, тип занятости, уровень опыта разработчика, год получения опыта работника, валюта (выплачиваемой зарплаты), местоположение сотрудника, мостоположение компании, размер компании, доля удалённой работы.\n",
|
|||
|
|
"\n",
|
|||
|
|
"*Связи между объектами:* Стаж сотрудника может быть использован в зарплатах и трудоустройстве по годам. Более высокий опыт и определённые должности коррелируют с большими зарплатами. Регионы и страы могут показать разницу в зарплатах. \n",
|
|||
|
|
"\n",
|
|||
|
|
"## Бизнес-цель\n",
|
|||
|
|
"1. Определение конкурентоспособности зарплат \n",
|
|||
|
|
"Цель: Сравнить зарплаты сотрудников по странам и должностям. \n",
|
|||
|
|
"Эффект же будет заключаться в том, что компании могут использовать эту информацию для привлечения и удержания талантов, предлагая конкрурентные зарплаты. \n",
|
|||
|
|
"2. Оценка зависимости между опытом и зарплатой \n",
|
|||
|
|
"Цель: Определить, как профессиональный опыт влияет на зарплаты в различных секторах. \n",
|
|||
|
|
"Эффект заключается в том, что компании могут лучше планировать бюджеты на повышение зарплат в зависимости от опыта сотрудников.\n",
|
|||
|
|
"\n",
|
|||
|
|
"### Техническая цель\n",
|
|||
|
|
"1. Конкурентоспособность зарплат \n",
|
|||
|
|
"Построить модель, которая будет предсказывать зарплаты сотрудников на основе атрибутов. На вход должны следовать такие признаки, как должность, опыт, страна компании. Целевой признак: зарплата.\n",
|
|||
|
|
"2. Оценка зависимости между опытом и зарплатой\n",
|
|||
|
|
"Создать модель и визуализировать влияние опыта на зарплату. На вход должны идти уровень опыта, должность, местоположение компании. Целевой признак: Зарплата"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 53,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1000x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1000x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Визуализация данных - ящик с усами. Выборка относительно сбалансирована, есть среднее смещение в среднюю сторону, медиана уравновешена\n",
|
|||
|
|
"plt.figure(figsize=(10, 6))\n",
|
|||
|
|
"sns.boxplot(x=df[\"salary_in_usd\"])\n",
|
|||
|
|
"plt.title(\"Box Plot для salary_in_usd\")\n",
|
|||
|
|
"plt.xlabel(\"salary_in_usd\")\n",
|
|||
|
|
"plt.show()\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Визуализируем отношение размера компании и зарплаты\n",
|
|||
|
|
"plt.figure(figsize=(10, 6))\n",
|
|||
|
|
"plt.scatter(df[\"salary_in_usd\"], df[\"experience_level\"])\n",
|
|||
|
|
"plt.xlabel(\"salary_in_usd\")\n",
|
|||
|
|
"plt.ylabel(\"experience_level\")\n",
|
|||
|
|
"plt.title(\"salary in usd vs experience_level\")\n",
|
|||
|
|
"plt.show()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"# Проблемы выбранный наборов данных \n",
|
|||
|
|
"**Зашумленность:** \n",
|
|||
|
|
"Возможное присутствие некорректных или неполных данных, например, ошибки в заполнении зарплат или местоположений.\n",
|
|||
|
|
"Необходима очистка данных от выбросов и ошибок. \n",
|
|||
|
|
"\n",
|
|||
|
|
"**Смещение:** \n",
|
|||
|
|
"Могут быть смещения в данных по странам или компаниям, например, если большая часть выборки состоит из сотрудников одной страны или отрасли. \n",
|
|||
|
|
"\n",
|
|||
|
|
"**Актуальность:** \n",
|
|||
|
|
"Если данные собраны за несколько лет, некоторые зарплаты могут быть неактуальны для текущего рынка, что требует дополнительной актуализации. \n",
|
|||
|
|
"\n",
|
|||
|
|
"**Выбросы:** \n",
|
|||
|
|
"Возможны экстремальные значения в зарплатах, которые необходимо либо исключить, либо обработать, чтобы не искажать модели. \n",
|
|||
|
|
"\n",
|
|||
|
|
"**Просачивание данных:** \n",
|
|||
|
|
"Если в данных есть признаки, которые могут напрямую предсказывать зарплату (например, если salary_currency или company_size напрямую коррелируют с зарплатой), это может привести к ошибкам в моделях."
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"### Борьба с выбросами"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 55,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1000x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1200x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"# Статистический анализ для определения выбросов\n",
|
|||
|
|
"Q1 = df[\"salary_in_usd\"].quantile(0.25)\n",
|
|||
|
|
"Q3 = df[\"salary_in_usd\"].quantile(0.75)\n",
|
|||
|
|
"IQR = Q3 - Q1\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Определение порога для выбросов\n",
|
|||
|
|
"threshold = 1.5 * IQR\n",
|
|||
|
|
"outliers = (df[\"salary_in_usd\"] < (Q1 - threshold)) | (\n",
|
|||
|
|
" df[\"salary_in_usd\"] > (Q3 + threshold)\n",
|
|||
|
|
")\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Обработка выбросов\n",
|
|||
|
|
"# В данном случае мы уберем выбросы\n",
|
|||
|
|
"median_salary = df[\"salary_in_usd\"].median()\n",
|
|||
|
|
"df.loc[outliers, \"salary_in_usd\"] = 0\n",
|
|||
|
|
"df1 = df[df.salary_in_usd != 0]\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Визуализация данных после обработки\n",
|
|||
|
|
"plt.figure(figsize=(10, 6))\n",
|
|||
|
|
"plt.scatter(df1[\"salary_in_usd\"], df1[\"experience_level\"])\n",
|
|||
|
|
"plt.xlabel(\"salary_in_usd\")\n",
|
|||
|
|
"plt.ylabel(\"experience_level\")\n",
|
|||
|
|
"plt.title(\"salary in usd vs experience_level\")\n",
|
|||
|
|
"plt.show()\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Визуализация данных после обработки\n",
|
|||
|
|
"plt.figure(figsize=(12, 6))\n",
|
|||
|
|
"plt.boxplot(df1[\"salary_in_usd\"])\n",
|
|||
|
|
"plt.xlabel(\"salary_in_usd\")\n",
|
|||
|
|
"plt.ylabel(\"experience_level\")\n",
|
|||
|
|
"plt.title(\"salary in usd vs experience_level\")\n",
|
|||
|
|
"plt.show()"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 57,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"name": "stdout",
|
|||
|
|
"output_type": "stream",
|
|||
|
|
"text": [
|
|||
|
|
"Размер обучающей выборки: 2253\n",
|
|||
|
|
"Размер контрольной выборки: 751\n",
|
|||
|
|
"Размер тестовой выборки: 751\n"
|
|||
|
|
]
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"from sklearn.model_selection import train_test_split\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Пример датасета с данными (X — признаки, y — целевой признак)\n",
|
|||
|
|
"X = df.drop(columns=[\"salary_in_usd\"]) # Признаки (все столбцы, кроме целевого признака 'salary')\n",
|
|||
|
|
"y = df[\"salary_in_usd\"] # Целевая переменная\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Разделяем данные на обучающую (60%), временную (40%) выборки\n",
|
|||
|
|
"X_train, X_temp, y_train, y_temp = train_test_split(X, y, test_size=0.4, random_state=42)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Временную выборку делим пополам на контрольную (20%) и тестовую (20%) выборки\n",
|
|||
|
|
"X_val, X_test, y_val, y_test = train_test_split(X_temp, y_temp, test_size=0.5, random_state=42)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Проверяем размер выборок\n",
|
|||
|
|
"print(f\"Размер обучающей выборки: {len(X_train)}\")\n",
|
|||
|
|
"print(f\"Размер контрольной выборки: {len(X_val)}\")\n",
|
|||
|
|
"print(f\"Размер тестовой выборки: {len(X_test)}\")"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"Теперь оценим сбалансированность, если нужно использовать методы аугментации данных, то применим. Анаализ проведём в категории 'salary_in_usd', в которой может потребоваться применение oversampling или undersampling."
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "code",
|
|||
|
|
"execution_count": 58,
|
|||
|
|
"metadata": {},
|
|||
|
|
"outputs": [
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "iVBORw0KGgoAAAANSUhEUgAAA14AAAIjCAYAAAATE8pZAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAACSP0lEQVR4nOzde3zO9f/H8ce1jR3MxowdYgi1OeaUY04p53NElErSOVHkWygdpKOS0MmhSEnoIOUs50MRGVGY0zAyNtuwfX5/vH+7MjYx17XPDs/77XbdfK7P9bk+13PXtV2u1/V5f15vh2VZFiIiIiIiIuI2HnYHEBERERERye9UeImIiIiIiLiZCi8RERERERE3U+ElIiIiIiLiZiq8RERERERE3EyFl4iIiIiIiJup8BIREREREXEzFV4iIiIiIiJupsJLREQkj0lMTGT//v38888/dkcRFzt9+jR79+4lMTHR7igi4mIqvERERPKAWbNmceutt1K0aFH8/f2JiIjg9ddftztWnpCQkMDYsWOd10+ePMn48ePtC3QBy7L48MMPqV+/Pn5+fgQEBFC+fHk+//xzu6OJiIs5LMuy7A4hIu4xZcoU7rvvPud1b29vIiIiuP322xk+fDghISE2phORK/Xss88yZswYOnXqRM+ePQkODsbhcHDDDTdQpkwZu+PleqmpqQQGBjJp0iSaNGnCW2+9xY4dO1iwYIHd0ejVqxdffvklffv2pX379gQGBuJwOKhevTolS5a0O56IuJCX3QFExP1GjRpF+fLlSU5OZuXKlUyYMIH58+ezbds2/Pz87I4nIpexfPlyxowZw+jRo3n22WftjpMneXp68uKLL3LPPfeQlpZGQEAAP/zwg92xmDZtGl9++SWff/45d911l91xRMTNdMRLJB9LP+K1YcMG6tSp41w/ePBg3n77bWbMmEGvXr1sTCgi/6VDhw6cOHGCVatW2R0lzztw4AD79+8nKiqKYsWK2R2HatWqUb16daZPn253FBHJATrHS6QAatGiBQB79uwB4MSJEzz99NNUq1YNf39/AgICaNOmDVu2bLnkvsnJybzwwgvccMMN+Pj4EBYWRteuXfnrr78A2Lt3Lw6HI8tLs2bNnPtatmwZDoeDL7/8kv/973+EhoZSpEgROnbsyP79+y957HXr1tG6dWsCAwPx8/OjadOmWX4YbdasWaaP/8ILL1yy7eeff07t2rXx9fUlKCiInj17Zvr4l/vZLpSWlsbYsWOpUqUKPj4+hISEMGDAgEsaIZQrV4727dtf8jiPPfbYJfvMLPsbb7xxyXMKkJKSwsiRI6lYsSLe3t6UKVOGIUOGkJKSkulzdaGHHnqISpUq4efnR1BQEC1atOCXX37JsM28efNo164d4eHheHt7U6FCBV566SVSU1MzbNesWTOqVq3Kpk2baNiwIb6+vpQvX56JEydm2C799yCry7333ntJzqxe3ylTpji3+eWXX+jevTsRERHO5+Gpp54iKSnJuc2999572cd2OBzs3bs3y+fr4vsXL16cZs2aXfKcZWXJkiXccsstFClShGLFitGpUyeio6MzbLN27VqqVq1Kz549CQoKwtfXl7p16zJ37lznNgkJCRQpUoQnn3zyksc4cOAAnp6ejB492pm5XLlyl2x38e/Yvn37eOSRR7jxxhvx9fWlRIkSdO/e/ZLnI/31W7ZsmXPdhg0buO222yhatChFihTJ9DmZMmUKDoeDjRs3OtfFxcVl+rvevn37TDNfyXvCCy+84Px7Kl26NA0aNMDLy4vQ0NBLcmcm/f7pl6JFi3LzzTdneP7h39/3rKS/f6T/jiYmJrJt2zbKlClDu3btCAgIyPK5Avj777/p3r07QUFB+Pn5Ub9+/UuO2l3Ne2qzZs0uee945ZVX8PDwYMaMGRnWX817r4hkTUMNRQqg9CKpRIkSgPkPfe7cuXTv3p3y5ctz5MgRJk2aRNOmTdm+fTvh4eGAOU+iffv2LF68mJ49e/Lkk09y+vRpFi5cyLZt26hQoYLzMXr16kXbtm0zPO6wYcMyzfPKK6/gcDgYOnQoR48eZezYsbRs2ZLNmzfj6+sLmA+obdq0oXbt2owcORIPDw8mT57sLAxuvvnmS/ZbunRp54fNhIQEHn744Uwfe/jw4fTo0YMHHniAY8eOMW7cOJo0acJvv/2W6bfiDz74ILfccgsA33zzDXPmzMlw+4ABA5xHG5944gn27NnD+++/z2+//caqVasoVKhQps/D1Th58qTzZ7tQWloaHTt2ZOXKlTz44INERUWxdetW3nnnHf78889LPixe7OzZs/Tp04fSpUtz4sQJJk2aROvWrYmOjiYiIgIwH5j9/f0ZNGgQ/v7+LFmyhBEjRnDq1CneeOONDPv7559/aNu2LT169KBXr1589dVXPPzwwxQuXJj7778/w7ZPPPEEdevWzbDugQceyDJrZGQkzz33HGA+sD/11FMZbp81axZnzpzh4YcfpkSJEqxfv55x48Zx4MABZs2aBZjXqmXLls773H333XTp0oWuXbs61/3XeTbBwcG88847gCly3n33Xdq2bcv+/fsve1Rl0aJFtGnThuuvv54XXniBpKQkxo0bR6NGjfj111+dhcbx48f58MMP8ff354knnqBkyZJ8/vnndO3alenTp9OrVy/8/f3p0qULX375JW+//Taenp7Ox/niiy+wLIvevXtf9ue42IYNG1i9ejU9e/akdOnS7N27lwkTJtCsWTO2b9+e5TDl3bt306xZM/z8/HjmmWfw8/Pjo48+omXLlixcuJAmTZpcVY6sZOc9Id1bb73FkSNHrurxPvvsM8D8rn3wwQd0796dbdu2ceONN2Yr//HjxwEYM2YMoaGhPPPMM/j4+GT6XB05coSGDRty5swZnnjiCUqUKMHUqVPp2LEjX3/9NV26dMmw7yt5T73Y5MmTef7553nrrbcyDHu8ludZRC5iiUi+NXnyZAuwFi1aZB07dszav3+/NXPmTKtEiRKWr6+vdeDAAcuyLCs5OdlKTU3NcN89e/ZY3t7e1qhRo5zrPv30Uwuw3n777UseKy0tzXk/wHrjjTcu2aZKlSpW06ZNndeXLl1qAdZ1111nnTp1yrn+q6++sgDr3Xffde67UqVKVqtWrZyPY1mWdebMGat8+fLWbbfddsljNWzY0Kpatarz+rFjxyzAGjlypHPd3r17LU9PT+uVV17JcN+tW7daXl5el6zftWuXBVhTp051rhs5cqR14VvpL7/8YgHW9OnTM9x3wYIFl6wvW7as1a5du0uyP/roo9bFb88XZx8yZIhVqlQpq3bt2hme088++8zy8PCwfvnllwz3nzhxogVYq1atuuTxLmf9+vUWYH399dfOdWfOnLlkuwEDBlh+fn5WcnKyc13Tpk0twHrrrbec61JSUqybbrrJKlWqlHX27FnLsv79PZg1a9Yl+y1SpIjVt2/fS9Y3atTIat68ufN6+u/d5MmTL5tz9OjRlsPhsPbt25fpz3vx8/xf+vbta5UtWzbDug8//NACrPXr11/2vunPw/Hjx53rtmzZYnl4eFj33HNPhkyAtWzZMue6M2fOWFFRUVZoaKjzefzpp58swPrxxx8zPE716tUz/I7cd999VkRExCV5Lv7ZM3v+1qxZYwHWtGnTnOvSX7+lS5dalmVZ3bp1szw9Pa1t27Y5t4mLi7NKlChh1a5d27ku/f1pw4YNznWZ/Z1almW1a9cuw/N8Ne8JF/+NHj161CpatKjVpk2bDLmzcvH9Lcuyfv75ZwuwvvrqK+e6pk2bWlWqVMlyPxf/jqZfL1y4sPXnn39meA4ufq4GDhxoARn+rk+fPm2VL1/eKleunPP9+0rfU9Pzpv9e/PDDD5aXl5c1ePDgDJmz894rIlnTUEORAqBly5aULFmSMmXK0LNnT/z9/ZkzZw7XXXcdYLodeniYt4PU1FSOHz+Ov78/N954I7/++qtzP7NnzyY4OJjHH3/8kse4eGjc1bjnnnsoWrSo8/odd9xBWFgY8+fPB2Dz5s3s2rWLu+66i+PHjxMXF0dcXByJiYnceuu
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1000x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1000x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "iVBORw0KGgoAAAANSUhEUgAAA2YAAAIjCAYAAABoNwiVAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAACelUlEQVR4nOzdd3gU5cLG4d+m90J6Qgihho6AICId6SKIIipSVLCAHrsfHsUu1iPYu2CnCCiodBDpRToJHUJLSAJJSG/z/ZGTPYSEFjZMynNf114mu7PvPBvWJE9m5n0thmEYiIiIiIiIiGnszA4gIiIiIiJS3amYiYiIiIiImEzFTERERERExGQqZiIiIiIiIiZTMRMRERERETGZipmIiIiIiIjJVMxERERERERMpmImIiIiIiJiMhUzERERERERk6mYiYiIiFRSL774IhaLpdh9tWvXZuTIkeYEEpEyUzETkcs2ZcoULBaL9ebi4kKDBg0YN24c8fHxZscTERERqXQczA4gIpXXyy+/TGRkJFlZWaxcuZJPPvmEP/74gx07duDm5mZ2PBGRamn37t3Y2elv7yKVjYqZiJRZnz59aNOmDQD33Xcffn5+/Oc//+HXX3/ljjvuMDmdiMjF5eXlUVBQgJOTk9lRbMbZ2dnsCCJSBvpziojYTLdu3QA4ePAgAKdOneLJJ5+kWbNmeHh44OXlRZ8+fdi6dWuJ52ZlZfHiiy/SoEEDXFxcCAkJ4ZZbbmH//v0AHDp0qNjpk+feunTpYh1r+fLlWCwWpk2bxrPPPktwcDDu7u4MGDCAI0eOlNj3unXr6N27N97e3ri5udG5c2dWrVpV6mvs0qVLqft/8cUXS2z7/fff07p1a1xdXalRowZDhw4tdf8Xem1nKygoYNKkSTRp0gQXFxeCgoK4//77OX36dLHtateuTf/+/UvsZ9y4cSXGLC3722+/XeJrCpCdnc0LL7xAvXr1cHZ2Jjw8nKeffprs7OxSv1Zne+CBB6hfvz5ubm7UqFGDbt268ffffxfb5tdff6Vfv36Ehobi7OxM3bp1eeWVV8jPzy+2XZcuXWjatCmbNm3i+uuvx9XVlcjISD799NNi2xW9D853K+0anPP9+06ZMsW6zd9//81tt91GrVq1rF+Hxx57jMzMTOs2I0eOvOC+LRYLhw4dOu/X62LPX758ebHtZ8yYYX2v+fv7M2zYMI4dO1Zi3JiYGIYMGUJAQACurq40bNiQf//73yW2q1279iXt988//6Rjx464u7vj6elJv3792Llz53lf19kOHDjAbbfdRo0aNXBzc+O6667j999/tz4eHx+Pg4MDL730Uonn7t69G4vFwocffmi9Lzk5mUcffZTw8HCcnZ2pV68eb775JgUFBdZtiv5fe+edd5g0aRJ169bF2dmZXbt2AfDBBx/QpEkT3Nzc8PX1pU2bNvz444/W5x8+fJiHHnqIhg0b4urqip+fH7fddluJf8ui071XrlzJI488QkBAAD4+Ptx///3k5OSQnJzM8OHD8fX1xdfXl6effhrDMErN+d577xEREYGrqyudO3dmx44dF/3annuNWVGeVatW8fjjjxMQEIC7uzuDBg0iISGh2HMLCgp48cUXCQ0Nxc3Nja5du7Jr1y5dtyZyFeiImYjYTFGJ8vPzAwp/8ZozZw633XYbkZGRxMfH89lnn9G5c2d27dpFaGgoAPn5+fTv358lS5YwdOhQ/vWvf3HmzBkWLVrEjh07qFu3rnUfd9xxB3379i223/Hjx5ea57XXXsNisfDMM89w8uRJJk2aRI8ePdiyZQuurq4ALF26lD59+tC6dWteeOEF7Ozs+Oabb6zFoW3btiXGrVmzJhMnTgQgLS2NBx98sNR9P//88wwZMoT77ruPhIQEPvjgAzp16sTmzZvx8fEp8ZwxY8bQsWNHAGbNmsXs2bOLPX7//fczZcoURo0axSOPPMLBgwf58MMP2bx5M6tWrcLR0bHUr8PlSE5Otr62sxUUFDBgwABWrlzJmDFjaNSoEdu3b+e9995jz549zJkz54Lj5uTkMGzYMGrWrMmpU6f47LPP6N27N9HR0dSqVQso/OXRw8ODxx9/HA8PD5YuXcqECRNITU3l7bffLjbe6dOn6du3L0OGDOGOO+5g+vTpPPjggzg5OXHPPfcU2/aRRx7h2muvLXbffffdd96sUVFR1rKSmJjIY489VuzxGTNmkJGRwYMPPoifnx/r16/ngw8+4OjRo8yYMQMo/Lfq0aOH9Tl33303gwYN4pZbbrHeFxAQcMGvmbOzM19++WWx+zZs2MD7779f7L6i98S1117LxIkTiY+PZ/LkyaxatarYe23btm107NgRR0dHxowZQ+3atdm/fz9z587ltddeK7H/jh07MmbMGACio6N5/fXXiz3+3XffMWLECHr16sWbb75JRkYGn3zyCTfccAObN2+mdu3a531t8fHxXH/99WRkZPDII4/g5+fH1KlTGTBgADNnzmTQoEEEBQXRuXNnpk+fzgsvvFDs+dOmTcPe3p7bbrsNgIyMDDp37syxY8e4//77qVWrFqtXr2b8+PGcOHGCSZMmFXv+N998Q1ZWFmPGjMHZ2ZkaNWrwxRdf8Mgjj3Drrbfyr3/9i6ysLLZt28a6deu48847rV//1atXM3ToUGrWrMmhQ4f45JNP6NKlC7t27SpxCvfDDz9McHAwL730EmvXruXzzz/Hx8eH1atXU6tWLV5//XX++OMP3n77bZo2bcrw4cOLPf/bb7/lzJkzjB07lqysLCZPnky3bt3Yvn07QUFB5/36ns/DDz+Mr68vL7zwAocOHWLSpEmMGzeOadOmWbcZP348b731FjfddBO9evVi69at9OrVi6ysrMven4hcJkNE5DJ98803BmAsXrzYSEhIMI4cOWL8/PPPhp+fn+Hq6mocPXrUMAzDyMrKMvLz84s99+DBg4azs7Px8ssvW+/7+uuvDcD4z3/+U2JfBQUF1ucBxttvv11imyZNmhidO3e2fr5s2TIDMMLCwozU1FTr/dOnTzcAY/Lkydax69evb/Tq1cu6H8MwjIyMDCMyMtK48cYbS+zr+uuvN5o2bWr9PCEhwQCMF154wXrfoUOHDHt7e+O1114r9tzt27cbDg4OJe7fu3evARhTp0613vfCCy8YZ3+L/vvvvw3A+OGHH4o9d/78+SXuj4iIMPr161ci+9ixY41zv+2fm/3pp582AgMDjdatWxf7mn733XeGnZ2d8ffffxd7/qeffmoAxqpVq0rs70LWr19vAMbMmTOt92VkZJTY7v777zfc3NyMrKws632dO3c2AOPdd9+13pednW20bNnSCAwMNHJycgzD+N/7YMaMGSXGdXd3N0aMGFHi/g4dOhhdu3a1fl70vvvmm28umHPixImGxWIxDh8+XOrrPffrfDEjRoww3N3dS9w/Y8YMAzCWLVtmGIZh5OTkGIGBgUbTpk2NzMxM63bz5s0zAGPChAnW+zp16mR4enqWyHj2e79IWFiYMWrUKOvnRV/Lov2eOXPG8PHxMUaPHl3seXFxcYa3t3eJ+8/16KOPGkCx99OZM2eMyMhIo3bt2tbvG5999pkBGNu3by/2/MaNGxvdunWzfv7KK68Y7u7uxp49e4pt93//93+Gvb29ERsbaxjG//49vby8jJMnTxbb9uabbzaaNGlywdyl/duvWbPGAIxvv/3Wel/R98hzv7e0b9/esFgsxgMPPGC9Ly8vz6hZs2ax/9+Kcp79/dQwDGPdunUGYDz22GPW+879XmEYhd8Dzn5/F+Xp0aNHsTyPPfaYYW9vbyQnJxuGUfjv5+DgYAwcOLDYeC+++KIBlPr/jIjYjk5lFJEy69GjBwEBAYSHhzN06FA8PDyYPXs2YWFhQOFf/IsuQM/PzycpKQkPDw8aNmzIP//8Yx3nl19+wd/fn4cffrjEPs499e5yDB8+HE9PT+vnt956KyEhIfzxxx8AbNmyhb1793LnnXeSlJREYmIiiYmJpKen0717d1asWFHsNCgoPOXSxcXlgvudNWsWBQUFDBkyxDpmYmIiwcHB1K9
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1000x600 with 1 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"name": "stdout",
|
|||
|
|
"output_type": "stream",
|
|||
|
|
"text": [
|
|||
|
|
"Количество записей в обучающей выборке до аугментации: 2253\n",
|
|||
|
|
"Количество записей в обучающей выборке после oversampling: 42746\n",
|
|||
|
|
"Количество записей в обучающей выборке после undersampling: 737\n",
|
|||
|
|
"\n",
|
|||
|
|
"Распределение salary_in_usd в обучающей выборке:\n",
|
|||
|
|
"salary_in_usd\n",
|
|||
|
|
"150000 58\n",
|
|||
|
|
"100000 58\n",
|
|||
|
|
"120000 53\n",
|
|||
|
|
"160000 52\n",
|
|||
|
|
"130000 48\n",
|
|||
|
|
" ..\n",
|
|||
|
|
"18000 1\n",
|
|||
|
|
"72212 1\n",
|
|||
|
|
"236600 1\n",
|
|||
|
|
"66192 1\n",
|
|||
|
|
"167580 1\n",
|
|||
|
|
"Name: count, Length: 737, dtype: int64\n",
|
|||
|
|
"\n",
|
|||
|
|
"Распределение salary_in_usd в контрольной выборке:\n",
|
|||
|
|
"salary_in_usd\n",
|
|||
|
|
"100000 24\n",
|
|||
|
|
"150000 21\n",
|
|||
|
|
"130000 16\n",
|
|||
|
|
"160000 15\n",
|
|||
|
|
"135000 15\n",
|
|||
|
|
" ..\n",
|
|||
|
|
"122000 1\n",
|
|||
|
|
"20000 1\n",
|
|||
|
|
"63040 1\n",
|
|||
|
|
"68318 1\n",
|
|||
|
|
"125976 1\n",
|
|||
|
|
"Name: count, Length: 365, dtype: int64\n",
|
|||
|
|
"\n",
|
|||
|
|
"Распределение salary_in_usd в тестовой выборке:\n",
|
|||
|
|
"salary_in_usd\n",
|
|||
|
|
"120000 24\n",
|
|||
|
|
"150000 19\n",
|
|||
|
|
"130000 18\n",
|
|||
|
|
"200000 17\n",
|
|||
|
|
"100000 17\n",
|
|||
|
|
" ..\n",
|
|||
|
|
"158677 1\n",
|
|||
|
|
"139500 1\n",
|
|||
|
|
"21669 1\n",
|
|||
|
|
"168400 1\n",
|
|||
|
|
"187000 1\n",
|
|||
|
|
"Name: count, Length: 361, dtype: int64\n",
|
|||
|
|
"\n",
|
|||
|
|
"\n"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"data": {
|
|||
|
|
"image/png": "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
|
|||
|
|
"text/plain": [
|
|||
|
|
"<Figure size 1000x800 with 2 Axes>"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
"metadata": {},
|
|||
|
|
"output_type": "display_data"
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"source": [
|
|||
|
|
"from imblearn.over_sampling import RandomOverSampler\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Визуализация распределения зарплат в обучающей выборке\n",
|
|||
|
|
"plt.figure(figsize=(10, 6))\n",
|
|||
|
|
"sns.histplot(y_train, bins=30, kde=True, color='blue')\n",
|
|||
|
|
"plt.title('Распределение зарплат в обучающей выборке')\n",
|
|||
|
|
"plt.xlabel('Зарплата в USD')\n",
|
|||
|
|
"plt.ylabel('Частота')\n",
|
|||
|
|
"plt.show()\n",
|
|||
|
|
"\n",
|
|||
|
|
"ros = RandomOverSampler(random_state=42)\n",
|
|||
|
|
"X_train_ros, y_train_ros = ros.fit_resample(X_train, y_train)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Визуализация распределения после oversampling\n",
|
|||
|
|
"plt.figure(figsize=(10, 6))\n",
|
|||
|
|
"sns.histplot(y_train_ros, bins=30, kde=True, color='green')\n",
|
|||
|
|
"plt.title('Распределение зарплат после oversampling')\n",
|
|||
|
|
"plt.xlabel('Зарплата в USD')\n",
|
|||
|
|
"plt.ylabel('Частота')\n",
|
|||
|
|
"plt.show()\n",
|
|||
|
|
"\n",
|
|||
|
|
"from imblearn.under_sampling import RandomUnderSampler\n",
|
|||
|
|
"\n",
|
|||
|
|
"rus = RandomUnderSampler(random_state=42)\n",
|
|||
|
|
"X_train_rus, y_train_rus = rus.fit_resample(X_train, y_train)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Визуализация распределения после oversampling\n",
|
|||
|
|
"plt.figure(figsize=(10, 6))\n",
|
|||
|
|
"sns.histplot(y_train_ros, bins=30, kde=True, color='green')\n",
|
|||
|
|
"plt.title('Распределение зарплат после oversampling')\n",
|
|||
|
|
"plt.xlabel('Зарплата в USD')\n",
|
|||
|
|
"plt.ylabel('Частота')\n",
|
|||
|
|
"plt.show()\n",
|
|||
|
|
"\n",
|
|||
|
|
"print(f\"Количество записей в обучающей выборке до аугментации: {len(y_train)}\")\n",
|
|||
|
|
"print(f\"Количество записей в обучающей выборке после oversampling: {len(y_train_ros)}\")\n",
|
|||
|
|
"print(f\"Количество записей в обучающей выборке после undersampling: {len(y_train_rus)}\")\n",
|
|||
|
|
"print()\n",
|
|||
|
|
"\n",
|
|||
|
|
"train_df, val_df = train_test_split(df, test_size=0.4, random_state=42)\n",
|
|||
|
|
"val_df, test_df = train_test_split(val_df, test_size=0.5, random_state=42)\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Функция для проверки распределения\n",
|
|||
|
|
"def check_balance(df, name):\n",
|
|||
|
|
" counts = df['salary_in_usd'].value_counts() # Обращение к столбцу 'salary_in_usd'\n",
|
|||
|
|
" print(f\"Распределение salary_in_usd в {name}:\")\n",
|
|||
|
|
" print(counts)\n",
|
|||
|
|
" print()\n",
|
|||
|
|
"\n",
|
|||
|
|
"check_balance(train_df, \"обучающей выборке\")\n",
|
|||
|
|
"check_balance(val_df, \"контрольной выборке\")\n",
|
|||
|
|
"check_balance(test_df, \"тестовой выборке\")\n",
|
|||
|
|
"print()\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Корреляция\n",
|
|||
|
|
"# Сначала отбираем только числовые столбцы\n",
|
|||
|
|
"numeric_df = df.select_dtypes(include='number')\n",
|
|||
|
|
"correlation_matrix = numeric_df.corr()\n",
|
|||
|
|
"\n",
|
|||
|
|
"# Визуализация корреляционной матрицы\n",
|
|||
|
|
"plt.figure(figsize=(10, 8))\n",
|
|||
|
|
"sns.heatmap(correlation_matrix, annot=True, fmt=\".2f\", cmap='coolwarm', square=True)\n",
|
|||
|
|
"plt.title(\"Корреляция между переменными\")\n",
|
|||
|
|
"plt.show()\n"
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"Таким образом, мы проанализировали 3 датасета. Посмотрели какие проблемы с ними могут быть связаны, возможные пути их решения. \n",
|
|||
|
|
"Разбили выборки, далее их сбалансировали, с помощью методов OverSampling и UnderSampling. Методы oversampling и undersampling лучше использовать для задач классификации, где целевая переменная — дискретные классы. Для регрессионных задач эти методы напрямую не применимы. Методы аугментации данных для регрессии, такие как SMOTE для регрессии, могут также помочь в увеличении количества обучающих данных и улучшении качества модели."
|
|||
|
|
]
|
|||
|
|
},
|
|||
|
|
{
|
|||
|
|
"cell_type": "markdown",
|
|||
|
|
"metadata": {},
|
|||
|
|
"source": [
|
|||
|
|
"Вроде бы всё... Вроде бы получилось :)"
|
|||
|
|
]
|
|||
|
|
}
|
|||
|
|
],
|
|||
|
|
"metadata": {
|
|||
|
|
"kernelspec": {
|
|||
|
|
"display_name": "mai",
|
|||
|
|
"language": "python",
|
|||
|
|
"name": "python3"
|
|||
|
|
},
|
|||
|
|
"language_info": {
|
|||
|
|
"codemirror_mode": {
|
|||
|
|
"name": "ipython",
|
|||
|
|
"version": 3
|
|||
|
|
},
|
|||
|
|
"file_extension": ".py",
|
|||
|
|
"mimetype": "text/x-python",
|
|||
|
|
"name": "python",
|
|||
|
|
"nbconvert_exporter": "python",
|
|||
|
|
"pygments_lexer": "ipython3",
|
|||
|
|
"version": "3.12.6"
|
|||
|
|
}
|
|||
|
|
},
|
|||
|
|
"nbformat": 4,
|
|||
|
|
"nbformat_minor": 2
|
|||
|
|
}
|