% Example 6.6
% 
% This one shows how to do agglomerative model-based clustering.
%
% Written by:
% -- 
% John L. Weatherwax                2009-03-24
% 
% 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' ); 

% Load up the data and put into a data matrix.
load iris
X = [setosa; versicolor; virginica];
% Do the agglomerative model-based clustering.
% This function is included in the EDA Toolbox
% and the MBC Toolbox.
Z = agmbclust(X);

% Construct a dendrogram.
dendrogram(Z);
title('Results for Iris Data - Agglomerative MBC')

% We can apply the silhouette procedure for this
% after we find a partition. Use 2/4 groups.
cind = cluster(Z,2);
figure; [S,H] = silhouette(X,cind);
title('Silhouette Plot - Agglomerative MBC: 2 clusters')
saveas( gcf, '../../WriteUp/Graphics/Chapter6/prob_6_9_iris_2_clusts', 'epsc' ); 

% after we find a partition. Use 2/4 groups.
cind = cluster(Z,4);
figure; [S,H] = silhouette(X,cind);
title('Silhouette Plot - Agglomerative MBC: 4 clusters')
saveas( gcf, '../../WriteUp/Graphics/Chapter6/prob_6_9_iris_4_clusts', 'epsc' ); 

% just to compare lets look at 3 groups:
cind = cluster(Z,3); 
figure; [S,H] = silhouette(X,cind);
title('Silhouette Plot - Agglomerative MBC: 3 clusters')