线性回归 Python实现

 import numpy as np
 import pylab

 def plot_data(data, b, m):
     x = data[:, 0]
     y = data[:, 1]
     y_predict = m*x + b
     pylab.plot(x, y_predict, 'k-')
     pylab.plot(x, y, 'o')
     pylab.show()

 def gradient(data, initial_b, initial_m, learning_rate, num_iter):
     b = initial_b
     m = initial_m
     x = data[:, 0]
     y = data[:, 1]
     n = float(len(data))
     for i in range(num_iter):
         b_gradient = -(1/n)*(y - m*x-b)
         b_gradient = np.sum(b_gradient, axis=0)
         m_gradient = -(1/n)*x*(y - m*x - b)
         m_gradient = np.sum(m_gradient)
         theta0 = b - (learning_rate*b_gradient)
         theta1 = m - (learning_rate*m_gradient)
         b = theta0
         m = theta1
         if i % 100 == 0:
             j = (np.sum((y - m*x - b)**2))/n
             print("参数b:{},参数m:{},损失值：{}".format(b, m, j))
     return [b, m]

 def linear_regression():
     # 导入数据
     data = np.loadtxt('data.csv', delimiter=',')

     # 线性回归参数设定
     learning_rate = 0.001
     initial_b = 0.0
     initial_m = 0.0
     num_iter = 1000

     b, m = gradient(data, initial_b, initial_m, learning_rate, num_iter)
     plot_data(data, b, m)

 if __name__ == '__main__':
     linear_regression()

测试结果：