%
% An exension of Example 6.8
%
% Written by:
% -- 
% John L. Weatherwax                2005-08-14
% 
% email: wax@alum.mit.edu
% 
% Please send comments and especially bug reports to the
% above email address.
% 
%-----

close all; drawnow; 
clc; 
clear; addpath( '../../Code/eda_data' ); addpath( '../../Code/eda_toolbox' ); addpath( '../Chapter1' ); 

% Uses the results from the previous example or Example 6.7, which 
% shows how to use the function for the whole MBC procedure.

load iris
data = [setosa;versicolor;virginica];
% Call the model-based clustering procedure with a maximum of 6 clusters.
[bics,bestmodel,allmodel,Z,clabs] = mbclust(data,6);

% Display the BIC curves.
plotbic(bics)

% Display the agglomerative part in the 
% One can use the clabs information along with gplotmatrix
% to do a grouped scatterplot matrix of the data. Do the groups seem
% reasonable?

% To extract model 9, *2* clusters, we use the following:
allmodel(9).clus(2)

% Extract the data:
pies = allmodel(9).clus(2).pies;
mus = allmodel(9).clus(2).mus;
vars = allmodel(9).clus(2).vars;

% The mixclass function returns group labels
% as well as a measure of the uncertainty in 
% classificaiton.
[clabs2,unc] = mixclass(data,pies,mus,vars);

% As before, we can use the silhouette procedure to assess the cluster results.
figure; 
[S, H] = silhouette(data,clabs2);
title('Iris Data - Model-Based Clustering')