第三章：简单线性模型Py...

话不多说，直接放上自己写的代码

• 
  
  
  from
  numpy
  import
  *
  
  
• 
  
  
  import
  random
  
  
• 
  
  
  """
  
  
  
• 
  
  
  dfdgffhhghghr
  
  
  
• 
  
  
  """
  
  
  
• 
  
  
  #一维线性方程的线性回归实现，通过求导获得代价函数的最值
  
  
  
• 
  
  def loadDataSet2():
  
  
• 
  
  
  ###数据导入 自己创一个拟合y=2x+1的数据
  
  
  
• 
  
  dataArr=[]
  
  
• 
  
  labelArr=[]
  
  
• 
  
  
  # myArr=[[0,1],[1,3],[2,4.8],[4,9.3],[-2,-2.8],[-3,-5.2],[1.5,5],[2.8,4],[-3.5,-4]]
  
  
  
• 
  
  
  for
  i
  in
  range(
  20
  ):
  
  
• 
  
  x=random.uniform(
  -5
  ,
  5
  )
  
  
• 
  
  dataArr.append(round(x,
  2
  ))
  
  
• 
  
  labelArr.append(
  2
  *x+
  1
  -random.uniform(
  -1
  ,
  1
  ))
  
  
• 
  
  
  
  
• 
  
  
  #for itemArr in myArr:
  
  
  
• 
  
  
  #dataArr.append(float(itemArr[0]))
  
  
  
• 
  
  
  # labelArr.append(float(itemArr[1]))
  
  
  
• 
  
  
  return
  dataArr,labelArr
  
  
• 
  
  
  
  
• 
  
  def calLeastSquareMethod(dataIn,labels):
  
  
• 
  
  length=len(dataIn)
  #数据个数
  
  
  
• 
  
  sumX=
  0.0
  ;sumXX=
  0.0
  
  
  
• 
  
  
  for
  data
  in
  dataIn:
  
  
• 
  
  sumX+=data
  
  
• 
  
  sumXX+=data*data
  
  
• 
  
  avgX=sumX/length
  
  
• 
  
  s=
  0.0
  
  
  
• 
  
  
  for
  i
  in
  range(length):
  
  
• 
  
  s+=labels*(dataIn-avgX)
  
  
• 
  
  weight=s/(sumXX-sumX*sumX/length)
  
  
• 
  
  
  
  
• 
  
  s2=
  0.0
  
  
  
• 
  
  
  for
  i
  in
  range(length):
  
  
• 
  
  s2+=labels-weight*dataIn
  
  
• 
  
  b=s2/length
  
  
• 
  
  
  return
  weight,b
  
  
• 
  
  
  
  
• 
  
  def plotBestFit2(weight,b):
  
  
• 
  
  
  ###画出最佳拟合直线
  
  
  
• 
  
  
  import
  matplotlib.pyplot
  as
  plt
  
  
• 
  
  print(weight,b)
  
  
• 
  
  
  # dataArr=array(dataMatrix) #矩阵转换为数组
  
  
  
• 
  
  n=len(dataArr)
  
  
• 
  
  xcord1=[];ycord1=[]
  
  
• 
  
  
  #xcord2=[];ycord2=[]
  
  
  
• 
  
  
  
  
• 
  
  
  # for i in range(n):
  
  
  
• 
  
  
  # if(int(labelMatrix)==1):
  
  
  
• 
  
  
  # xcord1.append(dataArr[i,1]);ycord1.append(dataArr[i,2])
  
  
  
• 
  
  
  # else:
  
  
  
• 
  
  
  # xcord2.append(dataArr[i,1]);ycord2.append(dataArr[i,2])
  
  
  
• 
  
  
  for
  i
  in
  range(n):
  
  
• 
  
  xcord1.append(dataArr);
  
  
• 
  
  ycord1.append(labelArr)
  
  
• 
  
  fig=plt.figure()
  
  
• 
  
  
  #在子图中画出样本点
  
  
  
• 
  
  ax=fig.add_subplot(
  111
  )
  
  
• 
  
  ax.scatter(xcord1,ycord1,s=
  30
  ,c=
  'red'
  ,marker=
  's'
  )
  
  
• 
  
  
  #ax.scatter(xcord2,ycord2,s=30,c='green')
  
  
  
• 
  
  
  
  
• 
  
  
  #画出拟合直线
  
  
  
• 
  
  x=arange(
  -5.0
  ,
  5.0
  ,
  0.1
  )
  
  
• 
  
  y=weight*x+b
  
  
• 
  
  
  
  
• 
  
  ax.plot(x,y)
  
  
• 
  
  plt.xlabel(
  'X'
  );plt.ylabel(
  'Y'
  )
  
  
• 
  
  plt.show()
  
  
• 
  
  
  
  
• 
  
  dataArr,labelArr=loadDataSet2()
  
  
• 
  
  weight,b=calLeastSquareMethod(dataArr,labelArr)
  
  
• 
  
  plotBestFit2(weight,b)
  
  

更多免费技术资料可关注：annalin1203