IIS_2023_1/tepechin_kirill_lab_1/lab1.py

75 lines
2.9 KiB
Python

import matplotlib.pyplot as plt
import numpy as np
from matplotlib.colors import ListedColormap
from sklearn.datasets import make_circles
from sklearn.linear_model import LinearRegression
from sklearn.linear_model import Ridge
from sklearn.metrics import mean_squared_error
from sklearn.model_selection import train_test_split
from sklearn.pipeline import make_pipeline
from sklearn.preprocessing import PolynomialFeatures, StandardScaler
rs = 42
X, y = make_circles(noise=0.2, factor=0.5, random_state=rs)
X = StandardScaler().fit_transform(X)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=.4, random_state=rs)
# Инициализируем модели
linear_model = LinearRegression()
poly_model = make_pipeline(PolynomialFeatures(4), LinearRegression())
ridge_model = make_pipeline(PolynomialFeatures(4), Ridge(alpha=1.0))
# Обучаем модели
linear_model.fit(X_train, y_train)
poly_model.fit(X_train, y_train)
ridge_model.fit(X_train, y_train)
# Предсказываем значения для тестового набора
y_pred_linear = linear_model.predict(X_test)
y_pred_poly = poly_model.predict(X_test)
y_pred_ridge = ridge_model.predict(X_test)
# Качество моделей
mse_linear = mean_squared_error(y_test, y_pred_linear)
mse_poly = mean_squared_error(y_test, y_pred_poly)
mse_ridge = mean_squared_error(y_test, y_pred_ridge)
models_data = [(linear_model, "Линейная", mse_linear), (poly_model, "Полиномиальная", mse_poly),
(ridge_model, "Греб. полиномиальная", mse_ridge)]
# Печатаем MSE
print(f"mse линейная: {mse_linear}")
print(f"mse полиномиальная: {mse_poly}")
print(f"mse гребневая полиномиальная: {mse_ridge}")
cm = plt.cm.RdBu
cm_bright = ListedColormap(['#FF0000', '#0000FF'])
fig, axs = plt.subplots(1, 3, figsize=(15, 7))
fig.suptitle('Сравнение регрессионных моделей')
# Функция отрисовки моделей
def draw_plot(model_data, i):
h = .02 # шаг регулярной сетки
x0_min, x0_max = X[:, 0].min() - .5, X[:, 0].max() + .5
x1_min, x1_max = X[:, 1].min() - .5, X[:, 1].max() + .5
xx0, xx1 = np.meshgrid(np.arange(x0_min, x0_max, h), np.arange(x1_min, x1_max, h))
Z = model_data[0].predict(np.c_[xx0.ravel(), xx1.ravel()]) # 3
Z = Z.reshape(xx0.shape)
axs[i].contourf(xx0, xx1, Z, cmap=cm, alpha=.8)
axs[i].set_xlim(xx0.min(), xx0.max())
axs[i].set_ylim(xx0.min(), xx1.max())
axs[i].set_xticks(())
axs[i].set_yticks(())
axs[i].set_title(model_data[1])
axs[i].text(xx0.max() - .3, xx1.min() + .3, ('%.2f' % model_data[2]).lstrip('0'),
size=15, horizontalalignment='right')
axs[i].scatter(X_test[:, 0], X_test[:, 1], c=y_test, cmap=cm_bright, alpha=0.6)
for i, model_data in enumerate(models_data):
draw_plot(model_data, i)
plt.savefig("plots.png")
plt.show()