Matlab实现线性回归和逻辑回归: Linear Regression & Logistic Regression

时间:2023-12-14 18:54:14

原文:http://blog.csdn.net/abcjennifer/article/details/7732417

本文为Maching Learning 栏目补充内容,为上几章中所提到单参数线性回归多参数线性回归和 逻辑回归的总结版。旨在帮助大家更好地理解回归,所以我在Matlab中分别对他们予以实现,在本文中由易到难地逐个介绍。

本讲内容:

Matlab 实现各种回归函数

=========================

基本模型

Y=θ01X1型---线性回归(直线拟合)

解决过拟合问题---Regularization

Y=1/(1+e^X)型---逻辑回归(sigmod 函数拟合)

=========================
 
第一部分:基本模型

 

在解决拟合问题的解决之前,我们首先回忆一下线性回归和逻辑回归的基本模型。

设待拟合参数 θn*1 和输入参数[ xm*n, ym*1 ] 。

对于各类拟合我们都要根据梯度下降的算法,给出两部分:

①   cost function(指出真实值y与拟合值h<hypothesis>之间的距离):给出cost function 的表达式,每次迭代保证cost function的量减小;给出梯度gradient,即cost function对每一个参数θ的求导结果。

function [ jVal,gradient ] = costFunction ( theta )

②   Gradient_descent(主函数):用来运行梯度下降算法,调用上面的cost function进行不断迭代,直到最大迭代次数达到给定标准或者cost function返回值不再减小。

function [optTheta,functionVal,exitFlag]=Gradient_descent( )

线性回归:拟合方程为hθ(x)=θ0x01x1+…+θnxn,当然也可以有xn的幂次方作为线性回归项(如Matlab实现线性回归和逻辑回归: Linear Regression & Logistic Regression),这与普通意义上的线性不同,而是类似多项式的概念。

其cost function 为:Matlab实现线性回归和逻辑回归: Linear Regression & Logistic Regression

逻辑回归:拟合方程为hθ(x)=1/(1+e^(θTx)),其cost function 为:Matlab实现线性回归和逻辑回归: Linear Regression & Logistic Regression

cost function对各θj的求导请自行求取,看第三章最后一图,或者参见后文代码。

后面,我们分别对几个模型方程进行拟合,给出代码,并用matlab中的fit函数进行验证。

 

 

第二部分:Y=θ01X1型---线性回归(直线拟合)
 

Matlab 线性拟合 & 非线性拟合中我们已经讲过如何用matlab自带函数fit进行直线和曲线的拟合,非常实用。而这里我们是进行ML课程的学习,因此研究如何利用前面讲到的梯度下降法(gradient descent)进行拟合。

cost function:
 function [ jVal,gradient ] = costFunction2( theta )
%COSTFUNCTION2 Summary of this function goes here
% linear regression -> y=theta0 + theta1*x
% parameter: x:m*n theta:n* y:m* (m=,n=)
% %Data
x=[;;;];
y=[1.1;2.2;2.7;3.8];
m=size(x,); hypothesis = h_func(x,theta);
delta = hypothesis - y;
jVal=sum(delta.^); gradient()=sum(delta)/m;
gradient()=sum(delta.*x)/m; end

其中,h_func是hypothesis的结果:

 function [res] = h_func(inputx,theta)
%H_FUNC Summary of this function goes here
%   Detailed explanation goes here %cost function
res= theta()+theta()*inputx;function [res] = h_func(inputx,theta)
end

Gradient_descent:

 function [optTheta,functionVal,exitFlag]=Gradient_descent( )
%GRADIENT_DESCENT Summary of this function goes here
% Detailed explanation goes here options = optimset('GradObj','on','MaxIter',);
initialTheta = zeros(,);
[optTheta,functionVal,exitFlag] = fminunc(@costFunction2,initialTheta,options); end

result:

 >> [optTheta,functionVal,exitFlag] = Gradient_descent()

 Local minimum found.

 Optimization completed because the size of the gradient is less than
the default value of the function tolerance. <stopping criteria details> optTheta = 0.3000
0.8600 functionVal = 0.0720 exitFlag =
即得y=0.3+0.86x;
验证:
 function [ parameter ] = checkcostfunc(  )
%CHECKC2 Summary of this function goes here
% check if the cost function works well
% check with the matlab fit function as standard %check cost function
x=[;;;];
y=[1.1;2.2;2.7;3.8]; EXPR= {'x',''};
p=fittype(EXPR);
parameter=fit(x,y,p); end

运行结果:

 >> checkcostfunc()

 ans = 

      Linear model:
ans(x) = a*x + b
Coefficients (with % confidence bounds):
a = 0.86 (0.4949, 1.225)
b = 0.3 (-0.6998, 1.3)

和我们的结果一样。下面画图:

 function PlotFunc( xstart,xend )
%PLOTFUNC Summary of this function goes here
% draw original data and the fitted %===================cost function ====linear regression
%original data
x1=[;;;];
y1=[1.1;2.2;2.7;3.8];
%plot(x1,y1,'ro-','MarkerSize',);
plot(x1,y1,'rx','MarkerSize',);
hold on; %fitted line - 拟合曲线
x_co=xstart:0.1:xend;
y_co=0.3+0.86*x_co;
%plot(x_co,y_co,'g');
plot(x_co,y_co); hold off;
end
Matlab实现线性回归和逻辑回归: Linear Regression & Logistic Regression
第三部分:解决过拟合问题---Regularization

过拟合问题解决方法我们已在第三章中讲过,利用Regularization的方法就是在cost function中加入关于θ的项,使得部分θ的值偏小,从而达到fit效果。

例如定义costfunction J(θ): jVal=(theta(1)-5)^2+(theta(2)-5)^2;

在每次迭代中,按照gradient descent的方法更新参数θ:θ(i)-=gradient(i),其中gradient(i)是J(θ)对θi求导的函数式,在此例中就有gradient(1)=2*(theta(1)-5), gradient(2)=2*(theta(2)-5)。

函数costFunction, 定义jVal=J(θ)和对两个θ的gradient:

 function [ jVal,gradient ] = costFunction( theta )
%COSTFUNCTION Summary of this function goes here
% Detailed explanation goes here jVal= (theta()-)^+(theta()-)^; gradient = zeros(,);
%code to compute derivative to theta
gradient() = * (theta()-);
gradient() = * (theta()-); end

Gradient_descent,进行参数优化

 function [optTheta,functionVal,exitFlag]=Gradient_descent( )
%GRADIENT_DESCENT Summary of this function goes here
% Detailed explanation goes here options = optimset('GradObj','on','MaxIter',);
initialTheta = zeros(,)
[optTheta,functionVal,exitFlag] = fminunc(@costFunction,initialTheta,options); end

matlab主窗口中调用,得到优化厚的参数(θ1,θ2)=(5,5)

  [optTheta,functionVal,exitFlag] = Gradient_descent()

 initialTheta =

 Local minimum found.

 Optimization completed because the size of the gradient is less than
the default value of the function tolerance. <stopping criteria details> optTheta = functionVal = exitFlag =

第四部分:Y=1/(1+e^X)型---逻辑回归(sigmod 函数拟合)

hypothesis function:

 function [res] = h_func(inputx,theta)

 %cost function
tmp=theta()+theta()*inputx;%m*
res=./(+exp(-tmp));%m* end

cost function:

 function [ jVal,gradient ] = costFunction3( theta )
%COSTFUNCTION3 Summary of this function goes here
% Logistic Regression x=[-; -; -; ; ; ; ];
y=[0.01; 0.05; 0.3; 0.45; 0.8; 1.1; 0.99];
m=size(x,); %hypothesis data
hypothesis = h_func(x,theta); %jVal-cost function & gradient updating
jVal=-sum(log(hypothesis+0.01).*y + (-y).*log(-hypothesis+0.01))/m;
gradient()=sum(hypothesis-y)/m; %reflect to theta1
gradient()=sum((hypothesis-y).*x)/m; %reflect to theta end

Gradient_descent:

 function [optTheta,functionVal,exitFlag]=Gradient_descent( )

  options = optimset('GradObj','on','MaxIter',);
initialTheta = [;];
[optTheta,functionVal,exitFlag] = fminunc(@costFunction3,initialTheta,options); end

运行结果:

  [optTheta,functionVal,exitFlag] = Gradient_descent()

 Local minimum found.

 Optimization completed because the size of the gradient is less than
the default value of the function tolerance. <stopping criteria details> optTheta = 0.3526
1.7573 functionVal = 0.2498 exitFlag =

画图验证:

Matlab实现线性回归和逻辑回归: Linear Regression & Logistic Regression
有朋友问,这里就补充一下logistic regression中gradient的推导:
Matlab实现线性回归和逻辑回归: Linear Regression & Logistic Regression
则有
Matlab实现线性回归和逻辑回归: Linear Regression & Logistic Regression
由于cost function
Matlab实现线性回归和逻辑回归: Linear Regression & Logistic Regression
可得
Matlab实现线性回归和逻辑回归: Linear Regression & Logistic Regression
所以gradient = -J'(theta) = (z-y)x