wolffd@0: %DEMNS1	Demonstrate Neuroscale for visualisation.
wolffd@0: %
wolffd@0: %	Description
wolffd@0: %	This script demonstrates the use of the Neuroscale algorithm for
wolffd@0: %	topographic projection and visualisation.  A data sample is generated
wolffd@0: %	from a mixture of two Gaussians in 4d space, and an RBF is trained
wolffd@0: %	with the stress error function to project the data into 2d.  The
wolffd@0: %	training data and a test sample are both plotted in this projection.
wolffd@0: %
wolffd@0: %	See also
wolffd@0: %	RBF, RBFTRAIN, RBFPRIOR
wolffd@0: %
wolffd@0: 
wolffd@0: %	Copyright (c) Ian T Nabney (1996-2001)
wolffd@0: 
wolffd@0: % Generate the data
wolffd@0: % Fix seeds for reproducible results
wolffd@0: rand('state', 420);
wolffd@0: randn('state', 420);
wolffd@0: 
wolffd@0: input_dim = 4;
wolffd@0: output_dim = 2;
wolffd@0: mix = gmm(input_dim, 2, 'spherical');
wolffd@0: mix.centres = [1 1 1 1; 0 0 0 0];
wolffd@0: mix.priors = [0.5 0.5];
wolffd@0: mix.covars = [0.1 0.1];
wolffd@0: 
wolffd@0: ndata = 60;
wolffd@0: [data, labels] = gmmsamp(mix, ndata);
wolffd@0: 
wolffd@0: clc
wolffd@0: disp('This demonstration illustrates the use of the Neuroscale model')
wolffd@0: disp('to perform a topographic projection of data.  We begin by generating')
wolffd@0: disp('60 data points from a mixture of two Gaussians in 4 dimensional space.')
wolffd@0: disp(' ')
wolffd@0: disp('Press any key to continue')
wolffd@0: pause
wolffd@0: 
wolffd@0: ncentres = 10;
wolffd@0: net = rbf(input_dim, ncentres, output_dim, 'tps', 'neuroscale');
wolffd@0: dstring = ['the Sammon mapping.  The model has ', num2str(ncentres), ...
wolffd@0:     ' centres, two outputs, and uses'];
wolffd@0: clc
wolffd@0: disp('The Neuroscale model is an RBF with a Stress error measure as used in')
wolffd@0: disp(dstring)
wolffd@0: disp('thin plate spline basis functions.')
wolffd@0: disp(' ')
wolffd@0: disp('It is trained using the shadow targets algorithm for at most 60 iterations.')
wolffd@0: disp(' ')
wolffd@0: disp('Press any key to continue')
wolffd@0: pause
wolffd@0: 
wolffd@0: % First row controls shadow targets, second row controls rbfsetbf
wolffd@0: options(1, :) = foptions;
wolffd@0: options(2, :) = foptions;
wolffd@0: options(1, 1) = 1;
wolffd@0: options(1, 2) = 1e-2;
wolffd@0: options(1, 3) = 1e-2;
wolffd@0: options(1, 6) = 1;    % Switch on PCA initialisation
wolffd@0: options(1, 14) = 60;
wolffd@0: options(2, 1) = -1;   % Switch off all warnings
wolffd@0: options(2, 5) = 1;
wolffd@0: options(2, 14) = 10;
wolffd@0: net2 = rbftrain(net, options, data);
wolffd@0: 
wolffd@0: disp(' ')
wolffd@0: disp('After training the model, we project the training data by a normal')
wolffd@0: disp('forward propagation through the RBF network.  Because there are two')
wolffd@0: disp('outputs, the results can be plotted and visualised.')
wolffd@0: disp(' ')
wolffd@0: disp('Press any key to continue')
wolffd@0: pause
wolffd@0: 
wolffd@0: % Plot the result
wolffd@0: y = rbffwd(net2, data);
wolffd@0: ClassSymbol1 = 'r.';
wolffd@0: ClassSymbol2 = 'b.';
wolffd@0: PointSize = 12;
wolffd@0: fh1 = figure;
wolffd@0: hold on;
wolffd@0: plot(y((labels==1),1),y(labels==1,2),ClassSymbol1, 'MarkerSize', PointSize)
wolffd@0: plot(y((labels>1),1),y(labels>1,2),ClassSymbol2, 'MarkerSize', PointSize)
wolffd@0: 
wolffd@0: disp(' ')
wolffd@0: disp('In this plot, the red dots denote the first class and the blue')
wolffd@0: disp('dots the second class.')
wolffd@0: disp(' ')
wolffd@0: disp('Press any key to continue.')
wolffd@0: disp(' ')
wolffd@0: pause
wolffd@0: 
wolffd@0: disp('We now generate a further 100 data points from the original distribution')
wolffd@0: disp('and plot their projection using star symbols.  Note that a Sammon')
wolffd@0: disp('mapping cannot be used to generalise to new data in this fashion.')
wolffd@0: 
wolffd@0: [test_data, test_labels] = gmmsamp(mix, 100);
wolffd@0: ytest = rbffwd(net2, test_data);
wolffd@0: ClassSymbol1 = 'ro';
wolffd@0: ClassSymbol2 = 'bo';
wolffd@0: % Circles are rather large symbols
wolffd@0: PointSize = 6;
wolffd@0: hold on
wolffd@0: plot(ytest((test_labels==1),1),ytest(test_labels==1,2), ...
wolffd@0:   ClassSymbol1, 'MarkerSize', PointSize)
wolffd@0: plot(ytest((test_labels>1),1),ytest(test_labels>1,2),...
wolffd@0:   ClassSymbol2, 'MarkerSize', PointSize)
wolffd@0: hold on
wolffd@0: legend('Class 1', 'Class 2', 'Test Class 1', 'Test Class 2')
wolffd@0: disp('Press any key to exit.')
wolffd@0: pause
wolffd@0: 
wolffd@0: close(fh1);
wolffd@0: clear all;
wolffd@0: