如何用Python实现常见机器学习算法-1

                1
                #
                 计算代价函数
                2
                def
                 computerCost(X,y,theta):

                3     m = len(y)
4     J = 0
5     J = (np.transpose(X*theta-y))*(X*theta-y)/(2*m) #计算代价J6return J

                 1
                #
                 梯度下降算法
                 2
                def
                 gradientDescent(X,y,theta,alpha,num_iters):

                 3     m = len(y)      
 4     n = len(theta)
 5     temp = np.matrix(np.zeros((n,num_iters)))   # 暂存每次迭代计算的theta，转化为矩阵形式 6     J_history = np.zeros((num_iters,1)) #记录每次迭代计算的代价值 7 8for i in range(num_iters):  # 遍历迭代次数     9         h = np.dot(X,theta)     # 计算内积，matrix可以直接乘10         temp[:,i] = theta - ((alpha/m)*(np.dot(np.transpose(X),h-y)))   #梯度的计算11         theta = temp[:,i]
12         J_history[i] = computerCost(X,y,theta)      #调用计算代价函数13print‘.‘,      
14return theta,J_history

                 1
                #
                 归一化feature
                 2
                def
                 featureNormaliza(X):

                 3     X_norm = np.array(X)            #将X转化为numpy数组对象，才可以进行矩阵的运算 4#定义所需变量 5     mu = np.zeros((1,X.shape[1]))   
 6     sigma = np.zeros((1,X.shape[1]))
 7 8     mu = np.mean(X_norm,0)          # 求每一列的平均值（0指定为列，1代表行） 9     sigma = np.std(X_norm,0)        # 求每一列的标准差10for i in range(X.shape[1]):     # 遍历列11         X_norm[:,i] = (X_norm[:,i]-mu[i])/sigma[i]  # 归一化1213return X_norm,mu,sigma

                 1
                #
                -*- coding: utf-8 -*-
                 2
                import
                 numpy as np

                 3
                from sklearn import linear_model
 4from sklearn.preprocessing import StandardScaler    #引入归一化的包 5 6def linearRegression():
 7print u"加载数据...\n" 8     data = loadtxtAndcsv_data("data.txt",",",np.float64)  #读取数据 9     X = np.array(data[:,0:-1],dtype=np.float64)      # X对应0到倒数第2列                  10     y = np.array(data[:,-1],dtype=np.float64)        # y对应最后一列  1112# 归一化操作13     scaler = StandardScaler()   
14    scaler.fit(X)
15     x_train = scaler.transform(X)
16     x_test = scaler.transform(np.array([1650,3]))
1718# 线性模型拟合19     model = linear_model.LinearRegression()
20    model.fit(x_train, y)
2122#预测结果23     result = model.predict(x_test)
24print model.coef_       # Coefficient of the features 决策函数中的特征系数25print model.intercept_  # 又名bias偏置,若设置为False，则为026print result            # 预测结果272829# 加载txt和csv文件30def loadtxtAndcsv_data(fileName,split,dataType):
31return np.loadtxt(fileName,delimiter=split,dtype=dataType)
3233# 加载npy文件34def loadnpy_data(fileName):
35return np.load(fileName)
363738if__name__ == "__main__":
39     linearRegression()