solved minor mistakes in BruteForceComparison

progs/testImageNetBinaryBruteForce.cpp

@@ -295,7 +295,7 @@ void inline trainGPSRMean(NICE::Vector & GPMeanRightPart, const double & noise,
     for (int run = 0; run < runsPerClassToAverageTraining; run++)
     {  
       NICE::Matrix innerMatrix;
-      innerMatrix.multiply(Kmn, Kmn, true /* tranpose first matrix*/, false /* transpose second matrix*/);
+      innerMatrix.multiply(Kmn, Kmn, false /* tranpose first matrix*/, true /* transpose second matrix*/);
       
       innerMatrix.addScaledMatrix( noise, Kmm );
       
@@ -322,7 +322,7 @@ void inline trainGPSRMean(NICE::Vector & GPMeanRightPart, const double & noise,
 }
 
 // GP subset of regressors
-void inline trainGPSRVar(NICE::Matrix choleskyMatrix, const double & noise, const NICE::Matrix & kernelMatrix, const int & nrOfExamplesPerClass, const int & classNumber, const int & runsPerClassToAverageTraining, const int & nrOfRegressors, std::vector<int> & indicesOfChosenExamples )
+void inline trainGPSRVar(NICE::Matrix & choleskyMatrix, const double & noise, const NICE::Matrix & kernelMatrix, const int & nrOfExamplesPerClass, const int & classNumber, const int & runsPerClassToAverageTraining, const int & nrOfRegressors, std::vector<int> & indicesOfChosenExamples )
 {
   std::vector<int> examplesToChoose;
   indicesOfChosenExamples.clear();
@@ -376,7 +376,7 @@ void inline trainGPSRVar(NICE::Matrix choleskyMatrix, const double & noise, cons
     for (int run = 0; run < runsPerClassToAverageTraining; run++)
     {  
       NICE::Matrix innerMatrix;
-      innerMatrix.multiply(Kmn, Kmn, true /* tranpose first matrix*/, false /* transpose second matrix*/);
+      innerMatrix.multiply(Kmn, Kmn, false /* tranpose first matrix*/, true /* transpose second matrix*/);
       
       innerMatrix.addScaledMatrix( noise, Kmm );
            
@@ -390,7 +390,7 @@ void inline trainGPSRVar(NICE::Matrix choleskyMatrix, const double & noise, cons
      }
  
     tTrainPrecise.stop(); 
-    std::cerr << "Precise time used for GPSRMean training class " << classNumber << ": " << tTrainPrecise.getLast()/(double)runsPerClassToAverageTraining << std::endl;    
+    std::cerr << "Precise time used for GPSRVar training class " << classNumber << ": " << tTrainPrecise.getLast()/(double)runsPerClassToAverageTraining << std::endl;    
 }
 
 KCMinimumEnclosingBall *trainSVDD( const double & noise, const NICE::Matrix kernelMatrix, const int & nrOfExamplesPerClass, const int & classNumber, const int & runsPerClassToAverageTraining )
@@ -535,13 +535,13 @@ void inline evaluateGPSRMean(const NICE::Vector & kernelVector,  const NICE::Vec
     {
       kernelVectorM[i] = kernelVector[indicesOfChosenExamples[i]];
     }
-    
+
     Timer tTestSingle;
     tTestSingle.start();
     
     for (int run = 0; run < runsPerClassToAverageTesting; run++)
     {
-      // \mean = \k_*^T \cdot K^{-1} \cdot y      
+      // \mean = \k_*^T \cdot K^{-1} \cdot y    
       mean = kernelVectorM.scalarProduct ( GPSRMeanRightPart );
     }
 
@@ -565,7 +565,7 @@ void inline evaluateGPSRVar(const NICE::Vector & kernelVector,  const NICE::Matr
     for (int i = 0; i < nrOfRegressors; i++)
     {
       kernelVectorM[i] = kernelVector[indicesOfChosenExamples[i]];
-    }    
+    } 
     
     Timer tTestSingle;
     tTestSingle.start();
@@ -573,6 +573,7 @@ void inline evaluateGPSRVar(const NICE::Vector & kernelVector,  const NICE::Matr
     for (int run = 0; run < runsPerClassToAverageTesting; run++)
     {
       NICE::Vector rightPart (nrOfRegressors,0.0);      
+      
       choleskySolveLargeScale ( choleskyMatrix, kernelVectorM, rightPart );
       
       uncertainty = noise*kernelVectorM.scalarProduct ( rightPart );
@@ -653,6 +654,7 @@ int main (int argc, char **argv)
   // share parameters among methods and classes?
   bool shareParameters = conf.gB("main" , "shareParameters", true);
   
+  //which methods do we want to use?
   bool GPMeanApprox = conf.gB( "main", "GPMeanApprox", false);
   bool GPVarApprox = conf.gB( "main", "GPVarApprox", false);  
   bool GPMean = conf.gB( "main", "GPMean", false);
@@ -661,6 +663,39 @@ int main (int argc, char **argv)
   bool GPSRVar = conf.gB( "main", "GPSRVar", false);  
   bool Parzen = conf.gB( "main", "Parzen", false);
   bool SVDD = conf.gB( "main", "SVDD", false);  
+  
+  if (GPMeanApprox)
+    std::cerr << "GPMeanApprox used" << std::endl;
+  else 
+    std::cerr << "GPMeanApprox not used" << std::endl;
+  if (GPVarApprox)
+    std::cerr << "GPVarApprox used" << std::endl;
+  else 
+    std::cerr << "GPVarApprox not used" << std::endl;
+  if (GPMean)
+    std::cerr << "GPMean used" << std::endl;
+  else 
+    std::cerr << "GPMean not used" << std::endl;
+  if (GPVar)
+    std::cerr << "GPVar used" << std::endl;
+  else 
+    std::cerr << "GPVar not used" << std::endl;
+  if (GPSRMean)
+    std::cerr << "GPSRMean used" << std::endl;
+  else 
+    std::cerr << "GPSRMean not used" << std::endl;
+  if (GPSRVar)
+    std::cerr << "GPSRVar used" << std::endl;
+  else 
+    std::cerr << "GPSRVar not used" << std::endl;
+  if (Parzen)
+    std::cerr << "Parzen used" << std::endl;
+  else 
+    std::cerr << "Parzen not used" << std::endl;
+  if (SVDD)
+    std::cerr << "SVDD used" << std::endl;
+  else 
+    std::cerr << "SVDD not used" << std::endl;
 
   
   // GP variance approximation  
@@ -1083,7 +1118,8 @@ int main (int argc, char **argv)
     OverallPerformanceSVDD += perfvalueSVDD;   
 
     // clean up memory used by SVDD
-    delete svdd;
+    if (SVDD)    
+      delete svdd;
   }
   
   OverallPerformanceGPVarApprox /= nrOfClassesToConcidere;