testdich.m

四种支持向量机用于函数拟合与模式识别的Matlab示例程序

function [X,I,alpha,theta] = testdich( dim, datanum )
% TESTDICH randomly creates test examples for dichotomy problem.
%  [X,I] = testgidich( dim, datanum )
%
% TESTDICH creats test data randomly for the algorithms which try 
%  to split two finite point sets by a hyperplane (e.g. Perceptron-like
%  algorithms). The created data are linearly separable. 
% 
% Input:
%  dim [1x1] dimension of generated data.
%  datanum [1x1] number of generated data (size of the test set).
%
% Output:
%  X [dim x datanum] contains generated test points.
%  I [1 x datanum] contains labnels of test points (1 or 2).
%  alpha [dim x 1] a normal vector of the hyperplane which correctly 
%      splits test data.
%  theta [1x1] a threshold of the hyperplane.
%
% See also
%   TESTANDR, PERCEPTR, KOZINEC, EKOZINEC, LINSVM
%

% Statistical Pattern Recognition Toolbox, Vojtech Franc, Vaclav Hlavac
% (c) Czech Technical University Prague, http://cmp.felk.cvut.cz
% Written Vojtech Franc (diploma thesis)
% Modifications
% 26-feb-2001 V.Franc

X = 2*rand(dim, datanum )-1;
I= zeros(1,datanum);

alpha = 2*rand(dim,1)-1;
theta = rand(1);


I=linclass(X,alpha,0);

X = X + theta/sum(alpha);