svmnlspex01fun.m

数据挖掘的新方法－支持向量机书中算法例子

%SVMnLSPex01FUN.m
%
%Object function and Constrained function for Two Dimension SVM Problem, Two Class and Separable Situation;
%
%Method from Christopher J. C. Burges:
%Object function and Constrained function for Two Dimension SVM Problem, Two Class and Separable Situation;
%
%Method from Thorsten Joachims:
%"Making Large-Scale SVM Learning Practical", function (4)/(5)/(6)
%	
%  Objective:   min "f(A)=-sum(A)+A'*Q*A/2" ,							function (4);
%  Subject to:  sum{A'*Y}=0 ,													function (5);
%					 0 <= ai <= C ,												function (6);
%The optimizing variables is "Lagrange Multipliers": A=[a1,a2,...,am],m is the number of total samples.
%

function [F,G]= SVMnLSPex01FUN(A)

%Samples matrix X with (m*n), m samples with dimention n;
	fid=fopen('a.dat','r');					%Get the samples
   m1=fread(fid,1,'float');				%The number of Position Set
   m2=fread(fid,1,'float');				%The number of Negative Set
   n=fread(fid,1,'float');					%Dimension of the samples
   m=m1+m2;										%
	for k=1:n
		X(1:m,k)=fread(fid,m,'float');
	end
   fclose(fid);
   
	Y=[ones(m1,1);-ones(m2,1)];				%Classfication value for both Positive/Negative samples
   
	%Object Function value F and Constrained function value G, G(i) is the value of i'th constrained function.
   
   Q = diag(Y)*K(X,X')*diag(Y);								%Component of matrix Q is Q(ij)=yi*yj*Xi*Xj';	 
   F = -sum(A)+ 0.5*(A'*Q*A);									%Object Function after Mapping;
	G = sum(A'*Y);													%Constraint Functions after Mapping; 
   
   
   function Kernel = K(U,E)									%Kernel function
      Kernel = (U*E+1).^4;