# -*- coding: utf-8 -*-

"""Untitled6.ipynb

Automatically generated by Colaboratory.

Original file is located at

https://colab.research.google.com/drive/19GtMkn-krgPsP1wxEEqNqxRwbD7FdLRZ

"""

import pandas as pd

import numpy as np

import matplotlib.pyplot as plt

import seaborn as sns

df = pd.read_excel("credits_100.xlsx")

print(df.head(10))

sns.pairplot(df, height=3, aspect=1.5, vars=['salary', 'credit', 'married'], diag_kind='kde', kind='reg')

print()

print('корреляция между числовыми признаками')

fields = ['age', 'salary', 'married', 'children', 'credit']

corr = df[fields].corr()

print(corr)

mask = np.zeros_like(corr, dtype=np.bool)

mask[np.triu_indices_from(mask)] = True

sns.heatmap(corr,mask=mask, cmap=sns.light_palette('grey'))

from mpl_toolkits.mplot3d import axes3d

fig = plt.figure(figsize=(10, 10))

ax = fig.add_subplot(111, projection='3d')

x1 = df['salary']

y1 = df['married']

z1 = df['credit']

ax.scatter(x1, y1, z1, marker='o', label='blue')

plt.figure(figsize=(16, 6))

df.boxplot(column=['salary', 'credit'], vert=False)

plt.show()

print('построение регрессионно модели')

from sklearn.model_selection import train_test_split

from sklearn.linear_model import LinearRegression

from sklearn import metrics

X = df[['credit','age', 'married', 'children']]

y = df['salary']

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.15, random_state=80)

print(X_test[:5])

print(y_test[:5])

#тренируем регрессию

regressor = LinearRegression()

regressor.fit(X_train, y_train)

print('коэф-ты регрессии')

print(regressor.coef_)

#расчетные значения у для X_test

y_pred = regressor.predict(X_test)

for y1, y2 in zip(y_test[:10], y_pred[:10]):

print(y1, y2)

E = [abs(y1 - y2) / ((y1+y2)/2) for y1, y2 in zip(y_test, y_pred)]

print(E)

avg_E = sum(E) / len(E)

print('оценка точности')

print(f'E среднее = {avg_E:.3f}')

accuracy = regressor.score(X_test, y_test)

print(accuracy)