Task02：线性回归

• 模型建立：线性回归原理、线性回归模型todo
• 学习策略：线性回归损失函数、代价函数、目标函数todo
• 算法求解：梯度下降法、牛顿法、拟牛顿法等todo
• 线性回归的评估指标todo
• sklearn参数详解todo
• 案例:

• 
   
  import numpy as np
  from sklearn.linear_model import LinearRegression
  import matplotlib.pyplot as plt
  
  #from sklearn.linear_model import ElasticNet
  #from sklearn.model_selection import train_test_split
  
  #生成随机数
  np.random.seed(1234)
  x = np.random.rand(500,3)
  #构建映射关系，模拟真实的数据待预测值,映射关系为y = 4.2 + 5.7*x1 + 10.8*x2，可自行设置值进行尝试
  y = x.dot(np.array([4.2,5.7,10.8]))
  
  print('X:',x.shape)
  
  print('Y:',y.shape)
  
  #
  # from sklearn.cross_decomposition import train_test_split
  # X_train,X_test,y_train,y_test=train_test_split(x,y,train_size=0.7)
  #
  # from sklearn import preprocessing
  # standard_X=preprocessing.StandardScaler()
  # X_train=standard_X.fit_transform(X_train)
  # X_test=standard_X.transform(X_test)
  #
  # standard_y=preprocessing.StandardScaler()
  # y_train=standard_y.fit_transform(y_train,reshape(-1,1))
  # y_test=standard_y.transform(y_test.reshape(-1,1))
  #
  #
  #
  #
  #
  # ElasticNet_clf =ElasticNet(alpha=0.1,l1_ratio=0.71)
  # ElasticNet_clf.fit(X_train,y_train.ravel())
  # ElasticNet_clf_score=ElasticNet_clf.score(X_test,y_test.ravel())
  
  #print('模型得分'，ElasticNet_clf_score)
  
  #print('权重'，ElasticNet_clf.coef_)
  #print('偏置'，ElasticNet_clf.intercept_)
  #print('迭代'，ElasticNet_clf.n_iter_)
  
  
  print('_______________________________________')
  
  #1、先尝试调用sklearn的线性回归模型训练数据
  
  # 调用模型
  lr = LinearRegression(fit_intercept=True)
  # 训练模型
  lr.fit(x,y)
  print("估计的参数值为：%s" %(lr.coef_))
  # 计算R平方
  print('R2:%s' %(lr.score(x,y)))
  # 任意设定变量，预测目标值
  x_test = np.array([2,4,5]).reshape(1,-1)
  y_hat = lr.predict(x_test)
  print("预测值为: %s" %(y_hat))
  
  
  
  
  #2、最小二乘法的矩阵求解
  
  class LR_LS():
      def __init__(self):
          self.w = None
      def fit(self, X, y):
          # 最小二乘法矩阵求解
          #============================= show me your code =======================
          self.w =np.linalg.inv(X.T.dot(X)).dot(X.T).dot(y)
          #============================= show me your code =======================
      def predict(self, X):
          # 用已经拟合的参数值预测新自变量
          #============================= show me your code =======================
          y_pred = X.dot(self.w)
          #============================= show me your code =======================
          return y_pred
  
  if __name__ == "__main__":
      lr_ls = LR_LS()
      lr_ls.fit(x,y)
      print("估计的参数值：%s" %(lr_ls.w))
      x_test = np.array([2,4,5]).reshape(1,-1)
      print("预测值为: %s" %(lr_ls.predict(x_test)))
  
  
  
  
  #3、梯度下降法
  
  class LR_GD():
      def __init__(self):
          self.w = None
  
      def fit(self, X, y, alpha=0.02, loss=1e-10):  # 设定步长为0.002,判断是否收敛的条件为1e-10
          y = y.reshape(-1, 1)  # 重塑y值的维度以便矩阵运算
          [m, d] = np.shape(X)  # 自变量的维度
          self.w = np.zeros((d))  # 将参数的初始值定为0
          tol = 1e5
          # ============================= show me your code =======================
          while tol > loss:
              h_f = X.dot(self.w).reshape(-1,1)
              theta = self.w + alpha*np.mean(X*(y - h_f),axis=0) #计算迭代的参数值
              tol = np.sum(np.abs(theta - self.w))
              self.w = theta
      # here
  
      # ============================= show me your code =======================
      def predict(self, X):
          # 用已经拟合的参数值预测新自变量
          y_pred = X.dot(self.w)
          return y_pred
  
  
  if __name__ == "__main__":
      lr_gd = LR_GD()
      lr_gd.fit(x, y)
      print("估计的参数值为：%s" % (lr_gd.w))
      x_test = np.array([2, 4, 5]).reshape(1, -1)
      print("预测值为：%s" % (lr_gd.predict(x_test)))
  
  
  

• 结果：
•  

• X: (500, 3)
  Y: (500,)
  _______________________________________
  估计的参数值为：[ 4.2  5.7 10.8]
  R2:1.0
  预测值为: [85.2]
  估计的参数值：[ 4.2  5.7 10.8]
  预测值为: [85.2]
  估计的参数值为：[ 4.20000001  5.70000003 10.79999997]
  预测值为：[85.19999995]