si_lab03

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import sklearn
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_absolute_error, mean_squared_error, mean_absolute_percentage_error

# x = np.arange(-3,3, 0.1).reshape((-1,1))
# y = np.tanh(x) + np.random.randn(*x.shape)*0.2
# ypred = LinearRegression().fit(x,y).predict(x)
# plt.scatter(x,y)
# plt.xlabel('x')
# plt.ylabel('y')
# plt.plot(x, ypred)
# plt.legend([ 'F(x) - aproksymująca',
#  'f(x) - aproksymowana zaszumiona'])
# plt.show()

bh_data = pd.read_excel('data.xlsx')
bh_arr = bh_data.values
X, y = bh_arr[:, :-1], bh_arr[:, -1]
# X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=221, shuffle=False)
# linReg = LinearRegression()
# linReg.fit(X_train, y_train)
# y_pred = linReg.predict(X_test)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, shuffle=True)

# linReg = LinearRegression()
# linReg.fit(X_train, y_train)
# y_pred = linReg.predict(X_test)
# mse = mean_squared_error(y_test, y_pred)
# mae = mean_absolute_error(y_test, y_pred)
# mape = mean_absolute_percentage_error(y_test, y_pred)
#
# minval = min(y_test.min(), y_pred.min())
# maxval = max(y_test.max(), y_pred.max())\
#
# plt.scatter(y_test, y_pred)
# plt.plot([minval, maxval], [minval, maxval])

outliers = np.abs((y_train - y_train.mean())/
 y_train.std())>3

nowe_dane = np.stack([X[:, 4]/X[:, 7],
 X[:, 4]/X[:, 5],
 X[:, 4]*X[:, 3],
X[:, 4]/X[:, -1]], axis=-1)
X_additional = np.concatenate([X, nowe_dane], axis=-1)

X_train_no_outliers = X_train[~outliers,:]
y_train_no_outliers = y_train[~outliers]
y_train_mean = y_train.copy()
y_train_mean[outliers] = y_train.mean()

linReg = LinearRegression()
linReg.fit(X_train, y_train_mean)

bh_cechy = list(bh_data.columns.values)

print(bh_cechy)

niezalezne_cechy = bh_cechy[:-1]

fig, ax = plt.subplots(1, 1)
x = np.arange(len(niezalezne_cechy))
wagi = linReg.coef_
ax.bar(x, wagi)
ax.set_xticks(x)
ax.set_xticklabels(niezalezne_cechy, rotation=90)

plt.boxplot(X_additional)
# plt.title("Medianowa wartosc mieszkania")
plt.xlabel('y_test')
plt.ylabel('y_pred')



plt.show()