# coding=utf-8
import numpy as np
import pandas as pd
import math
import matplotlib.pyplot as plt
from sklearn import datasets, linear_model
from sklearn.metrics import mean_squared_error, r2_score
from sklearn import preprocessing
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import PolynomialFeatures


data = pd.read_csv('ticdata20001.txt', sep=" ", header=None, engine='python')
data.columns = ["1", "2", "3", "4", "5", "6", "7", "8", "9", "10", "11", "12", "13"]

y = data.loc[:,  ["1", "2", "3", "4", "5", "6", "7", "8", "9", "10", "11", "12", "13"]]
x = data.loc[:, ["1", "2", "3", "4", "5", "6", "7", "8", "9", "10", "11", "12", "13"] ]
all_rows = len(data.index)
train_data = int(math.floor(all_rows * 0.5))
test_data = int(math.floor(all_rows * 0.5))

x_train = x[:train_data]
x_test = x[train_data:]

y_train = y.iloc[:train_data]
y_test = y.iloc[train_data:]

x_train = preprocessing.scale(x_train)
x_test = preprocessing.scale(x_test)

regr = linear_model.LinearRegression()
# передаем тренировочную выборку в качестве параметра
regr.fit(x_train, y_train)
# предсказываем значение по модели
y_pred = regr.predict(x_test)


print('Coefficients: \n', regr.coef_)
print("Mean squared error: %.100f" % mean_squared_error(y_test, y_pred))
print('Variance score: %.2f' % r2_score(y_test, y_pred))

print('\nPOLINOM\n')

# степень полинома в массиве
d = [1, 2, 3, 4, 5]
varScr = []
for i in d:
    polynomial_features = PolynomialFeatures(degree=i, include_bias=False)
# построение полимиальной функции
    pipeline = Pipeline([("polynomial_features", polynomial_features), ("linear_regression", regr)])
# расчет модели
    pipeline.fit(np.array(x_train), np.array(y_train))
# получаем предсказанные значения на основе тестовой выборки
    y_pred = pipeline.predict(np.array(x_test))
# сравнение тестовых и предсказанных значений
    r2 = r2_score(y_test, y_pred)
    varScr.append(r2)
    print("Mean squared error: %.100f" % mean_squared_error(y_test, y_pred))
    print('Variance score: %.2f' % r2)


fig = plt.figure()
plt.plot(d, varScr)
plt.xlabel('Degree polinome', fontsize=15)
plt.ylabel('Variance score', fontsize=15)
plt.show()

print('\nLASSO\n')
alphas = [0.0001, 0.001, 0.01, 0.1, 1]
varScr = []
for i in alphas:
    regr = linear_model.Lasso(alpha=i)
    regr.fit(x_train, y_train)
    y_pred = regr.predict(x_test)

    r2 = r2_score(y_test, y_pred)
    varScr.append(r2)
    # The mean squared error
    print("Mean squared error: %.100f"
          % mean_squared_error(y_test, y_pred))
    # Explained variance score: 1 is perfect prediction
    print('Variance score: %.2f' % r2)

fig = plt.figure()
plt.plot(alphas, varScr)
plt.xlabel('Regularization coefficient (d)', fontsize=15)
plt.ylabel('Variance score (varScr)', fontsize=15)
plt.show()