%
% Examples 9.7
% 
% Written by:
% -- 
% John L. Weatherwax                2008-02-20
% 
% email: wax@alum.mit.edu
% 
% Please send comments and especially bug reports to the
% above email address.
% 
%-----

clear all; 
close all; 
clc; 

addpath('../../Code/CSTool');
load iris
% This loads up three matrices: setosa, versicolor and virginica.
% 
% We will use the classes SETOSA AND VIRGINICA.
% 
% Note that the priors are equal, so the decision is
% based only on the class-conditional probabilities.
ncc = 0;
setosa(:,3:4)=[];
virginica(:,3:4)=[];
[nset,d]=size(setosa);
[nvir,d]=size(virginica);
n = nvir + nset;

% Loop first through all of the patterns corresponding
% to setosa. Here correct classification 
% is obtained if pxgver > pxgvir;
muvir  = mean(virginica);  
covvir = cov(virginica);
% These will be the same for this part.
for i = 1:nset
  % Get the test point and the training set
  settrain = setosa;
  % This is the testing point.
  x = settrain(i,:);
  % Delete from training set.
  % The result is the training set.
  settrain(i,:)=[];	
  muset = mean(settrain);
  covset = cov(settrain);
  pxgset=csevalnorm(x,muset,covset);
  pxgvir=csevalnorm(x,muvir,covvir);
  if pxgset > pxgvir % then we correctly classified it
    ncc = ncc+1;
  end
end

% Loop through all of the patterns of virginica notes.
% Here correct classification is obtained when
% pxgvir > pxxgver
muset = mean(setosa);
covset = cov(setosa);
% These remain the same for the following.
for i = 1:nvir
  % Get the test point and training set
  virtrain = virginica;
  x = virtrain(i,:);
  virtrain(i,:)=[];
  muvir = mean(virtrain);
  covvir = cov(virtrain);
  pxgset = csevalnorm(x,muset,covset);
  pxgvir = csevalnorm(x,muvir,covvir);
  if pxgvir > pxgset % then we correctly classified it
    ncc = ncc+1;
  end
end

pcc = ncc/n;
fprintf('pcc = %10.6f\n',pcc);