网络编程
位置:首页>> 网络编程>> Python编程>> Python编程实现线性回归和批量梯度下降法代码实例

Python编程实现线性回归和批量梯度下降法代码实例

作者:Key_Ky  发布时间:2021-10-13 07:33:27 

标签:线性回归,梯度下降,python

通过学习斯坦福公开课的线性规划和梯度下降,参考他人代码自己做了测试,写了个类以后有时间再去扩展,代码注释以后再加,作业好多:


import numpy as np
import matplotlib.pyplot as plt
import random

class dataMinning:
 datasets = []
 labelsets = []

addressD = '' #Data folder
 addressL = '' #Label folder

npDatasets = np.zeros(1)
 npLabelsets = np.zeros(1)

cost = []
 numIterations = 0
 alpha = 0
 theta = np.ones(2)
 #pCols = 0
 #dRows = 0
 def __init__(self,addressD,addressL,theta,numIterations,alpha,datasets=None):
   if datasets is None:
     self.datasets = []
   else:
     self.datasets = datasets
   self.addressD = addressD
   self.addressL = addressL
   self.theta = theta
   self.numIterations = numIterations
   self.alpha = alpha

def readFrom(self):
   fd = open(self.addressD,'r')
   for line in fd:
     tmp = line[:-1].split()
     self.datasets.append([int(i) for i in tmp])
   fd.close()
   self.npDatasets = np.array(self.datasets)

fl = open(self.addressL,'r')
   for line in fl:
     tmp = line[:-1].split()
     self.labelsets.append([int(i) for i in tmp])
   fl.close()

tm = []
   for item in self.labelsets:
     tm = tm + item
   self.npLabelsets = np.array(tm)

def genData(self,numPoints,bias,variance):
   self.genx = np.zeros(shape = (numPoints,2))
   self.geny = np.zeros(shape = numPoints)

for i in range(0,numPoints):
     self.genx[i][0] = 1
     self.genx[i][1] = i
     self.geny[i] = (i + bias) + random.uniform(0,1) * variance

def gradientDescent(self):
   xTrans = self.genx.transpose() #
   i = 0
   while i < self.numIterations:
     hypothesis = np.dot(self.genx,self.theta)
     loss = hypothesis - self.geny
     #record the cost
     self.cost.append(np.sum(loss ** 2))
     #calculate the gradient
     gradient = np.dot(xTrans,loss)
     #updata, gradientDescent
     self.theta = self.theta - self.alpha * gradient
     i = i + 1

def show(self):
   print 'yes'

if __name__ == "__main__":
 c = dataMinning('c:\\city.txt','c:\\st.txt',np.ones(2),100000,0.000005)
 c.genData(100,25,10)
 c.gradientDescent()
 cx = range(len(c.cost))
 plt.figure(1)
 plt.plot(cx,c.cost)
 plt.ylim(0,25000)
 plt.figure(2)
 plt.plot(c.genx[:,1],c.geny,'b.')
 x = np.arange(0,100,0.1)
 y = x * c.theta[1] + c.theta[0]
 plt.plot(x,y)
 plt.margins(0.2)
 plt.show()

Python编程实现线性回归和批量梯度下降法代码实例

图1. 迭代过程中的误差cost

Python编程实现线性回归和批量梯度下降法代码实例

图2. 数据散点图和解直线

总结

Python算法输出1-9数组形成的结果为100的所有运算式

python中实现k-means聚类算法详解

Python编程实现粒子群算法(PSO)详解

如有不足之处,欢迎留言指出。感谢朋友们对本站的支持!

来源:https://www.cnblogs.com/Key-Ky/archive/2013/12/10/3468290.html

0
投稿

猜你喜欢

手机版 网络编程 asp之家 www.aspxhome.com