386 KiB
386 KiB
Вариант: Список людей.
In [63]:
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.preprocessing import LabelEncoder
from sklearn import metrics
from imblearn.over_sampling import RandomOverSampler
from imblearn.under_sampling import RandomUnderSampler
from sklearn.preprocessing import StandardScaler, OneHotEncoder
from sklearn.metrics import ConfusionMatrixDisplay
from sklearn.compose import ColumnTransformer
from sklearn.pipeline import Pipeline
from sklearn.impute import SimpleImputer
from sklearn.linear_model import LinearRegression, LogisticRegression
from sklearn.ensemble import RandomForestRegressor, GradientBoostingRegressor, RandomForestClassifier, GradientBoostingClassifier
from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.metrics import (
precision_score, recall_score, accuracy_score, roc_auc_score, f1_score,
matthews_corrcoef, cohen_kappa_score, confusion_matrix
)
from sklearn.metrics import mean_squared_error, r2_score, mean_absolute_error
import numpy as np
import featuretools as ft
from sklearn.metrics import accuracy_score, classification_report
# Функция для применения oversampling
def apply_oversampling(X, y):
oversampler = RandomOverSampler(random_state=42)
X_resampled, y_resampled = oversampler.fit_resample(X, y)
return X_resampled, y_resampled
# Функция для применения undersampling
def apply_undersampling(X, y):
undersampler = RandomUnderSampler(random_state=42)
X_resampled, y_resampled = undersampler.fit_resample(X, y)
return X_resampled, y_resampled
def split_stratified_into_train_val_test(
df_input,
stratify_colname="y",
frac_train=0.6,
frac_val=0.15,
frac_test=0.25,
random_state=None,
):
"""
Splits a Pandas dataframe into three subsets (train, val, and test)
following fractional ratios provided by the user, where each subset is
stratified by the values in a specific column (that is, each subset has
the same relative frequency of the values in the column). It performs this
splitting by running train_test_split() twice.
Parameters
----------
df_input : Pandas dataframe
Input dataframe to be split.
stratify_colname : str
The name of the column that will be used for stratification. Usually
this column would be for the label.
frac_train : float
frac_val : float
frac_test : float
The ratios with which the dataframe will be split into train, val, and
test data. The values should be expressed as float fractions and should
sum to 1.0.
random_state : int, None, or RandomStateInstance
Value to be passed to train_test_split().
Returns
-------
df_train, df_val, df_test :
Dataframes containing the three splits.
"""
if frac_train + frac_val + frac_test != 1.0:
raise ValueError(
"fractions %f, %f, %f do not add up to 1.0"
% (frac_train, frac_val, frac_test)
)
if stratify_colname not in df_input.columns:
raise ValueError("%s is not a column in the dataframe" % (stratify_colname))
X = df_input # Contains all columns.
y = df_input[
[stratify_colname]
] # Dataframe of just the column on which to stratify.
# Split original dataframe into train and temp dataframes.
df_train, df_temp, y_train, y_temp = train_test_split(
X, y, stratify=y, test_size=(1.0 - frac_train), random_state=random_state
)
# Split the temp dataframe into val and test dataframes.
relative_frac_test = frac_test / (frac_val + frac_test)
df_val, df_test, y_val, y_test = train_test_split(
df_temp,
y_temp,
stratify=y_temp,
test_size=relative_frac_test,
random_state=random_state,
)
assert len(df_input) == len(df_train) + len(df_val) + len(df_test)
return df_train, df_val, df_test
df = pd.read_csv("../data/age.csv", nrows=10000)
df.info()
Как бизнес-цели выделим следующие 2 варианта: 1) GameDev. Создание игры про конкретного персонажа, живущего в конкретном временном промежутке в конкретной стране. 2) Исследование зависимости длительности жизни от страны проживания.
Поскольку именно эти бизнес-цели были выбраны в предыдущей лабораторной работе, будем их использовать. Но возникает проблема с 1 целью: её невозможно использовать для задачи классификации. Заменим ее на классификацию людей по возрастным группам, что может быть полезно для рекламных целей.
Выполним подготовку данных
In [46]:
df.fillna({"Gender": "NaN", "Country": "NaN", "Occupation" : "NaN", "Manner of death" : "NaN"}, inplace=True)
df = df.dropna()
df['Country'] = df['Country'].str.split('; ')
df = df.explode('Country')
data = df.copy()
value_counts = data["Country"].value_counts()
rare = value_counts[value_counts < 100].index
data = data[~data["Country"].isin(rare)]
data.drop(data[~data['Gender'].isin(['Male', 'Female'])].index, inplace=True)
data1 = pd.get_dummies(data, columns=['Gender', 'Country', 'Occupation'], drop_first=True)
Определить достижимый уровень качества модели для каждой задачи. На основе имеющихся данных уровень качества моделей не будет высоким, поскольку все таки длительность жизни лишь примерная и точно ее угадать невозможно.
Выберем ориентиры для наших 2х задач: 1)Регрессии - средний возраст человека 2)Классификации - аиболее часто встречающаяся возрастная группа
Построим конвейер.
In [33]:
print(data.columns)
In [37]:
X_reg = data1.drop(['Id', 'Name', 'Age of death', 'Short description', 'Manner of death'], axis=1)
y_reg = data1['Age of death']
# Разделение данных
X_train_reg, X_test_reg, y_train_reg, y_test_reg = train_test_split(X_reg, y_reg, test_size=0.2, random_state=42)
# Выбор моделей для регрессии
models_reg = {
'Linear Regression': LinearRegression(),
'Random Forest Regressor': RandomForestRegressor(random_state=42),
'Gradient Boosting Regressor': GradientBoostingRegressor(random_state=42)
}
# Создание конвейера для регрессии
pipelines_reg = {}
for name, model in models_reg.items():
pipelines_reg[name] = Pipeline([
('scaler', StandardScaler()),
('model', model)
])
# Определение сетки гиперпараметров для регрессии
param_grids_reg = {
'Linear Regression': {},
'Random Forest Regressor': {
'model__n_estimators': [100, 200, 300],
'model__max_depth': [None, 10, 20, 30]
},
'Gradient Boosting Regressor': {
'model__n_estimators': [100, 200, 300],
'model__learning_rate': [0.01, 0.1, 0.2],
'model__max_depth': [3, 5, 7]
}
}
# Настройка гиперпараметров для регрессии
best_models_reg = {}
for name, pipeline in pipelines_reg.items():
grid_search = GridSearchCV(pipeline, param_grids_reg[name], cv=5, scoring='neg_mean_squared_error')
grid_search.fit(X_train_reg, y_train_reg)
best_models_reg[name] = grid_search.best_estimator_
print(f'Best parameters for {name}: {grid_search.best_params_}')
# Обучение моделей и оценка качества
for model_name in best_models_reg.keys():
print(f"Model: {model_name}")
model = best_models_reg[model_name]["model"]
model_pipeline = Pipeline([("scaler", StandardScaler()), ("model", model)])
model_pipeline = model_pipeline.fit(X_train_reg, y_train_reg)
y_train_predict = model_pipeline.predict(X_train_reg)
y_test_predict = model_pipeline.predict(X_test_reg)
best_models_reg[model_name]["pipeline"] = model_pipeline
best_models_reg[model_name]["preds_train"] = y_train_predict
best_models_reg[model_name]["preds_test"] = y_test_predict
best_models_reg[model_name]["MSE_train"] = mean_squared_error(y_train_reg, y_train_predict)
best_models_reg[model_name]["MSE_test"] = mean_squared_error(y_test_reg, y_test_predict)
best_models_reg[model_name]["R2_train"] = r2_score(y_train_reg, y_train_predict)
best_models_reg[model_name]["R2_test"] = r2_score(y_test_reg, y_test_predict)
best_models_reg[model_name]["MAE_train"] = mean_absolute_error(y_train_reg, y_train_predict)
best_models_reg[model_name]["MAE_test"] = mean_absolute_error(y_test_reg, y_test_predict)
In [50]:
data2 = data.drop(['Short description', 'Manner of death', 'Gender', 'Country', 'Occupation'], axis=1)
In [64]:
# Создание возрастных групп
bins = [0, 18, 30, 50, 70, 100]
labels = ['0-18', '19-30', '31-50', '51-70', '71+']
data['Age Group'] = pd.cut(data['Age of death'], bins=bins, labels=labels)
# Выбор признаков и целевой переменной для классификации
X_class = data2.drop(['Id', 'Name', 'Age of death', 'Age Group'], axis=1)
y_class = data['Age Group']
print(X_class.columns)
# Разделение данных
X_train_class, X_test_class, y_train_class, y_test_class = train_test_split(X_class, y_class, test_size=0.2, random_state=42)
# Выбор моделей для классификации
models_class = {
'Logistic Regression': LogisticRegression(random_state=42, max_iter=5000, solver='liblinear'),
'Random Forest Classifier': RandomForestClassifier(random_state=42),
'Gradient Boosting Classifier': GradientBoostingClassifier(random_state=42)
}
# Создание конвейера для классификации
pipelines_class = {}
for name, model in models_class.items():
pipelines_class[name] = Pipeline([
('scaler', StandardScaler()),
('model', model)
])
# Определение сетки гиперпараметров для классификации
'''
param_grids_class = {
'Logistic Regression': {
'model__C': [0.1, 1, 10],
'model__solver': ['lbfgs', 'liblinear']
},
'Random Forest Classifier': {
'model__n_estimators': [100, 200, 300],
'model__max_depth': [None, 10, 20, 30]
},
'Gradient Boosting Classifier': {
'model__n_estimators': [100, 200, 300],
'model__learning_rate': [0.01, 0.1, 0.2],
'model__max_depth': [3, 5, 7]
}
}'''
# Убрал определение параметров поскольку уже был предподсчет данных, но вылетела ошибка. Сохранил лучшие параметры
param_grids_class = {
'Logistic Regression': {
'model__C': [10],
'model__solver': ['lbfgs']
},
'Random Forest Classifier': {
'model__n_estimators': [200],
'model__max_depth': [ 30]
},
'Gradient Boosting Classifier': {
'model__n_estimators': [200],
'model__learning_rate': [0.1],
'model__max_depth': [7]
}
}
# Настройка гиперпараметров для классификации
best_models_class = {}
for name, pipeline in pipelines_class.items():
grid_search = GridSearchCV(pipeline, param_grids_class[name], cv=5, scoring='accuracy', n_jobs=-1)
grid_search.fit(X_train_class, y_train_class)
best_models_class[name] = {"model": grid_search.best_estimator_}
print(f'Best parameters for {name}: {grid_search.best_params_}')
# Обучение моделей и оценка качества
for model_name in best_models_class.keys():
print(f"Model: {model_name}")
model = best_models_class[model_name]["model"]
model_pipeline = Pipeline([("scaler", StandardScaler()), ("model", model)])
model_pipeline = model_pipeline.fit(X_train_class, y_train_class)
y_train_predict = model_pipeline.predict(X_train_class)
y_test_probs = model_pipeline.predict_proba(X_test_class)
y_test_predict = model_pipeline.predict(X_test_class)
best_models_class[model_name]["pipeline"] = model_pipeline
best_models_class[model_name]["probs"] = y_test_probs
best_models_class[model_name]["preds"] = y_test_predict
best_models_class[model_name]["Precision_train"] = precision_score(y_train_class, y_train_predict, average='weighted')
best_models_class[model_name]["Precision_test"] = precision_score(y_test_class, y_test_predict, average='weighted')
best_models_class[model_name]["Recall_train"] = recall_score(y_train_class, y_train_predict, average='weighted')
best_models_class[model_name]["Recall_test"] = recall_score(y_test_class, y_test_predict, average='weighted')
best_models_class[model_name]["Accuracy_train"] = accuracy_score(y_train_class, y_train_predict)
best_models_class[model_name]["Accuracy_test"] = accuracy_score(y_test_class, y_test_predict)
best_models_class[model_name]["ROC_AUC_test"] = roc_auc_score(y_test_class, y_test_probs, multi_class='ovr')
best_models_class[model_name]["F1_train"] = f1_score(y_train_class, y_train_predict, average='weighted')
best_models_class[model_name]["F1_test"] = f1_score(y_test_class, y_test_predict, average='weighted')
best_models_class[model_name]["MCC_test"] = matthews_corrcoef(y_test_class, y_test_predict)
best_models_class[model_name]["Cohen_kappa_test"] = cohen_kappa_score(y_test_class, y_test_predict)
best_models_class[model_name]["Confusion_matrix"] = confusion_matrix(y_test_class, y_test_predict)
In [67]:
num_models = len(best_models_class)
fig, ax = plt.subplots(num_models, 1, figsize=(12, 10), sharex=False, sharey=False)
for index, key in enumerate(best_models_class.keys()):
c_matrix = best_models_class[key]["Confusion_matrix"]
disp = ConfusionMatrixDisplay(
confusion_matrix=c_matrix, display_labels=["0-18", "19-30", "31-50", "51-70", "71+"]
).plot(ax=ax.flat[index])
disp.ax_.set_title(key)
plt.subplots_adjust(top=1, bottom=0, hspace=0.4, wspace=0.1)
plt.show()
In [78]:
_, ax = plt.subplots(3, 2, figsize=(12, 10), sharex=False, sharey=False)
ax = ax.flatten()
for index, (name, model) in enumerate(best_models_reg.items()):
y_pred_reg = model.predict(X_test_reg)
# График фактических значений против предсказанных значений
ax[index * 2].scatter(y_test_reg, y_pred_reg, alpha=0.5)
ax[index * 2].plot([min(y_test_reg), max(y_test_reg)], [min(y_test_reg), max(y_test_reg)], color='red', linestyle='--')
ax[index * 2].set_xlabel('Actual Values')
ax[index * 2].set_ylabel('Predicted Values')
ax[index * 2].set_title(f'{name}: Actual vs Predicted')
# График остатков
residuals = y_test_reg - y_pred_reg
ax[index * 2 + 1].scatter(y_pred_reg, residuals, alpha=0.5)
ax[index * 2 + 1].axhline(y=0, color='red', linestyle='--')
ax[index * 2 + 1].set_xlabel('Predicted Values')
ax[index * 2 + 1].set_ylabel('Residuals')
ax[index * 2 + 1].set_title(f'{name}: Residuals vs Predicted')
plt.subplots_adjust(top=1, bottom=0, hspace=0.4, wspace=0.1)
plt.show()
