04.03.2015 Views

Algoritmo das Projeções Sucessivas Para Seleção de ... - PPGQ

Algoritmo das Projeções Sucessivas Para Seleção de ... - PPGQ

Algoritmo das Projeções Sucessivas Para Seleção de ... - PPGQ

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Anexos<br />

Programas Auxiliares do SPA-LDA<br />

• multilda.m<br />

• mahal2.m<br />

1) multilda.m<br />

function [class,errors] = multilda(sample,training,group,Group_Test)<br />

%MULTILDA Multi-Class Linear discriminant analysis.<br />

% [class,errors] = multilda(sample,training,group,Group_Test) classifies each row<br />

% of the data in SAMPLE into one of the values of the vector<br />

% GROUP. GROUP contains integers from one to the number of<br />

% groups in the training set, which is the matrix, TRAINING.<br />

%<br />

% SAMPLE and TRAINING must have the same number of columns.<br />

% TRAINING and GROUP must have the same number of rows.<br />

% CLASS is a vector with the same number of rows as SAMPLE.<br />

%<br />

% The biased Maximum-Likelihood estimate for the covariance matrix is employed: cov(X,1)<br />

%<br />

% B.A. Jones 2-05-95<br />

% Copyright 1993-2000 The MathWorks, Inc.<br />

% $Revision: 2.9 $ $Date: 2000/05/26 17:28:36 $<br />

[gr,gc] = size(group);<br />

if min(gr,gc) ~= 1<br />

error('Requires the third argument to be a vector.');<br />

end<br />

if gc ~= 1,<br />

group = group(:);<br />

gr = gc;<br />

end<br />

if any(group - round(group)) | any(group < 1)<br />

error('The third input argument must be positive integers.');<br />

end<br />

maxg = max(group);<br />

[tr,tc] = size(training);<br />

121

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!