from sklearn.linear_model import LinearRegression, SGDRegressor, Ridge, LogisticRegression
from sklearn.datasets import load_boston
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.metrics import mean_squared_error
from sklearn.externals import joblib
from sklearn.metrics import r2_score
from sklearn.neural_network import MLPRegressor

import pandas as pd
import numpy as np

lb = load_boston()
# train_test_split(train_data,train_target,test_size=0.3,random_state=5)
#train_data：待划分样本数据
#train_target：待划分样本数据的结果（标签）
#test_size：测试数据占样本数据的比例，若整数则样本数量
#random_state：设置随机数种子，保证每次都是同一个随机数。若为0或不填，则每次得到数据都不一样
#train_test_split()函数是用来随机划分样本数据为训练集和测试集的，当然也可以人为的切片划分
x_train, x_test, y_train, y_test = train_test_split(lb.data, lb.target, test_size=0.2)


# 为数据增加一个维度,相当于把[1, 5, 10] 变成 [[1, 5, 10],]
y_train = y_train.reshape(-1, 1)
y_test = y_test.reshape(-1, 1)

# 进行标准化
std_x = StandardScaler()
x_train = std_x.fit_transform(x_train)
x_test = std_x.transform(x_test)

std_y = StandardScaler()
y_train = std_y.fit_transform(y_train)
y_test = std_y.transform(y_test)

# 正规方程预测
#最小二乘法线性回归
lr = LinearRegression()
#fit_transform方法是fit和transform的结合，fit_transform(X_train) 意思是找出X_train的均值和标准差，并应用在X_train上
lr.fit(x_train, y_train)
print("r2 score of Linear regression is",r2_score(y_test,lr.predict(x_test)))

#岭回归
from sklearn.linear_model import RidgeCV
#岭回归模型
cv = RidgeCV(alphas=np.logspace(-3, 2, 100))
cv.fit (x_train , y_train)
print("r2 score of Linear regression is",r2_score(y_test,cv.predict(x_test)))

#梯度下降 用于判断使用凸loss函数（convex loss function）的分类器
sgd = SGDRegressor()
#一个数组X（其size为[n_samples, n_features]）：保存着训练样本；一个数组Y：保存着训练样本的target值（class label）：
sgd.fit(x_train, y_train)
print("r2 score of Linear regression is",r2_score(y_test,sgd.predict(x_test)))


from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense


#基准NN
#使用标准化后的数据
seq = Sequential()
#构建神经网络模型
#input_dim来隐含的指定输入数据shape
seq.add(Dense(64, activation=‘relu‘,input_dim=lb.data.shape[1]))
seq.add(Dense(64, activation=‘relu‘))
seq.add(Dense(1, activation=‘relu‘))
seq.compile(optimizer=‘rmsprop‘, loss=‘mse‘, metrics=[‘mae‘])
seq.fit(x_train, y_train,  epochs=300, batch_size = 16, shuffle = False)
score = seq.evaluate(x_test, y_test,batch_size=16) #loss value & metrics values
print("score:",score)
print(‘r2 score:‘,r2_score(y_test, seq.predict(x_test)))