ANBC_Model.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.
*/

#define GRT_DLL_EXPORTS
#include "ANBC_Model.h"

GRT_BEGIN_NAMESPACE

bool ANBC_Model::train( UINT classLabel, const MatrixFloat &trainingData, const VectorFloat &weightsVector ){
 
 //Check to make sure the column sizes match
 if( trainingData.getNumCols() != weightsVector.size() ){
  N = 0;
  return false;
 }
 
 UINT M = trainingData.getNumRows();
 N = trainingData.getNumCols();
 this->classLabel = classLabel;
 
 //Update the weights buffer
 weights = weightsVector;
 
 //Resize the buffers
 mu.resize( N );
 sigma.resize( N );
 
 //Calculate the mean for each dimension
 for(UINT j=0; j<N; j++){
  mu[j] = 0.0;
  
  for(UINT i=0; i<M; i++){
   mu[j] += trainingData[i][j];
  }
  
  mu[j] /= Float(M);
 }
 
 //Calculate the sample standard deviation
 for(UINT j=0; j<N; j++){
  sigma[j] = 0.0;
  
  for(UINT i=0; i<M; i++){
   sigma[j] += grt_sqr( trainingData[i][j]-mu[j] );
  }
  sigma[j] = grt_sqrt( sigma[j]/Float(M-1) );
  
  if( sigma[j] == 0 ){
   sigma[j] = 0.1; //Set the sigma to a small value so sigma is not zero!
  }
 }
 
 //Now compute the threshold
 Float meanPrediction = 0.0;
 VectorFloat predictions(M);
 VectorFloat testData(N);
 for(UINT i=0; i<M; i++){
  //Test the ith training example
  for(UINT j=0; j<N; j++) {
   testData[j] = trainingData[i][j];
  }
  
  predictions[i] = predict( testData );
  meanPrediction += predictions[i];
 }
 
 //Calculate the mean prediction value
 meanPrediction /= Float(M);
 
 //Calculate the standard deviation
 Float stdDev = 0.0;
 for(UINT i=0; i<M; i++) {
  stdDev += grt_sqr( predictions[i]-meanPrediction );
 }
 stdDev = grt_sqrt( stdDev / (Float(M)-1.0) );
 
 threshold = meanPrediction-(stdDev*gamma);
 
 //Update the training mu and sigma values so the threshold value can be dynamically computed at a later stage
 trainingMu = meanPrediction;
 trainingSigma = stdDev;
 
 return true;
}

Float ANBC_Model::predict( const VectorFloat &x ){
 Float prediction = 0.0;
 for(UINT j=0; j<N; j++){
  if(weights[j]>0)
  prediction += grt_log(gauss(x[j],mu[j],sigma[j]) * weights[j]);
 }
 return prediction;
}

Float ANBC_Model::predictUnnormed( const VectorFloat &x ){
 Float prediction = 0.0;
 for(UINT j=0; j<N; j++){
  if(weights[j]>0)
  prediction += grt_log( unnormedGauss(x[j], mu[j], sigma[j]) * weights[j] );
 }
 return prediction;
}

inline Float ANBC_Model::gauss(const Float x,const Float mu,const Float sigma){
 return ( 1.0/(sigma*sqrt(TWO_PI)) ) * exp( - ( ((x-mu)*(x-mu))/(2*(sigma*sigma)) ) );
}

inline Float ANBC_Model::unnormedGauss(const Float x,const Float mu,const Float sigma){
 return exp( - ( ((x-mu)*(x-mu))/(2*(sigma*sigma)) ) );
}

void ANBC_Model::recomputeThresholdValue(const Float gamma){
 this->gamma = gamma;
 threshold = trainingMu-(trainingSigma*gamma);
}

GRT_END_NAMESPACE