ClusteringModulesExamples/GaussianMixtureModelsExample/GaussianMixtureModelsExample.cpp

/*
 GRT MIT License
 Copyright (c) <2012> <Nicholas Gillian, Media Lab, MIT>
 
 Permission is hereby granted, free of charge, to any person obtaining a copy of this software
 and associated documentation files (the "Software"), to deal in the Software without restriction,
 including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense,
 and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so,
 subject to the following conditions:
 
 The above copyright notice and this permission notice shall be included in all copies or substantial
 portions of the Software.
 
 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT
 LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
 IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
 WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
 SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 */

/*
 GRT Gaussian Mixture Models Example
 This examples demonstrates how to use the GaussianMixtureModules module for unsupervised clustering.
 */

//You might need to set the specific path of the GRT header relative to your project
#include <GRT/GRT.h>
using namespace GRT;
using namespace std;

//This is a helper function to generate some dummy training data
MatrixFloat generateTrainingData(const UINT numClusters,const UINT numDimensions=3,const UINT numSamples=1000);

int main (int argc, const char * argv[])
{
 //Setup the stats that will be used to generate the training data
 UINT numClusters = 3;
 UINT numDimensions = 3;
 UINT numSamples = 10000;
 
 //Generate some training data to train the GMM algorithm
 MatrixFloat trainingData = generateTrainingData( numClusters, numDimensions, numSamples );
 
 //Setup the GMM instance
 GaussianMixtureModels gmm;
 gmm.setNumClusters( numClusters );
 gmm.setMinChange( 1.0e-10 );
 gmm.setMinNumEpochs( 10 );
 gmm.setMaxNumEpochs( 10000 );
 
 cout << "Training model...\n";
 
 //Train the GMM model using the training data
 if( !gmm.train( trainingData ) ){
 cout << "Failed to train model!\n";
 return EXIT_FAILURE;
 }
 
 cout << "GMM Trained in " << gmm.getNumTrainingIterationsToConverge() << " iterations.\n\n";
 
 //Save the model to a file
 if( !gmm.save( "GMM.grt" ) ){
 cout << "Failed to save model to file!\n";
 return EXIT_FAILURE;
 }
 
 //Load the model back from a file
 if( !gmm.load( "GMM.grt" ) ){
 cout << "Failed to load model from file!\n";
 return EXIT_FAILURE;
 }
 
 //Get the clusters from the GMM instance and print them
 //Note that the clusters may not be in the same order as they appear in the original model used to generate the training data
 MatrixFloat mu = gmm.getMu();
 Vector< MatrixFloat > sigma = gmm.getSigma();
 
 cout << "Mu:\n";
 for(unsigned int k=0; k<mu.getNumRows(); k++){
 for(unsigned int n=0; n<mu.getNumCols(); n++){
 cout << mu[k][n] << "\t";
 }cout << endl;
 }
 cout << endl;
 
 cout << "Sigma:\n";
 for(unsigned int k=0; k<sigma.size(); k++){
 cout << "Cluster: " << k+1 << endl;
 for(unsigned int m=0; m<sigma[k].getNumRows(); m++){
 for(unsigned int n=0; n<sigma[k].getNumCols(); n++){
 cout << sigma[k][m][n] << "\t";
 }
 cout << endl;
 }
 }
 cout << endl;
 
 return EXIT_SUCCESS;
}

MatrixFloat generateTrainingData(const UINT numClusters,const UINT numDimensions,const UINT numSamples){
 
 Random rand;
 MatrixFloat data(numSamples,numDimensions);
 
 //Generate a random model
 MatrixFloat mu(numClusters,numDimensions);
 MatrixFloat sigma(numClusters,numDimensions);
 for(UINT k=0; k<numClusters; k++){
 for(UINT n=0; n<numDimensions; n++){
 mu[k][n] = rand.getRandomNumberUniform(-10,10);
 sigma[k][n] = rand.getRandomNumberUniform(0.5,2.0);
 }
 }
 
 cout << "Model Mu:\n";
 for(UINT k=0; k<numClusters; k++){
 for(UINT n=0; n<numDimensions; n++){
 cout << mu[k][n] << "\t";
 }
 cout << endl;
 }
 cout << endl;
 
 cout << "Model Sigma:\n";
 for(UINT k=0; k<numClusters; k++){
 for(UINT n=0; n<numDimensions; n++){
 cout << sigma[k][n] << "\t";
 }
 cout << endl;
 }
 cout << endl;
 
 //Use the model to generate some training data
 for(UINT i=0; i<numSamples; i++){
 UINT randIndex = rand.getRandomNumberInt(0, numClusters);
 for(UINT n=0; n<numDimensions; n++){
 data[i][n] = mu[randIndex][n] + rand.getRandomNumberGauss(0,sigma[randIndex][n]);
 }
 }
 
 return data;
}