void NeuralClassifier::TEST_DATA(char* test_file, char* model, int n_hu){
Allocator *allocator = new Allocator;
int num_of_features = 9;
int my_digit = 5;
cout<<"Testing.......... ."<<endl;
DiskXFile model_file(model, "r");
InitClassifier();
cout<<"Reading data .......... ."<<endl;
DataSet *test_data =new(allocator) MatDataSet(test_file,num_of_features,1,false,-1,false);
/*Local variables & Initialization*/
Timer timer;
accuracy=0;
//ofstream fout("test_stats.txt");
cout<<"Forwarding samples .......... ."<<endl;
//DataSet *test_data, Machine* machine,int ExpectedDigit, char* StatFile
TestExamplesForDigit(test_data,mlp, my_digit,"test_stat.txt");
delete(allocator);
}
LinearSvmModel::LinearSvmModel(const std::string &filename)
{
if(filesystem::exists(filename) == false)
{
log_error() << "Could not find the linear SVM model file " << filename << std::endl;
throw std::invalid_argument("Could not find linear SVM model file");
}
log_info() << "Parsing linear svm model " << filename << std::endl;
// open and parse the file
std::ifstream model_file(filename.c_str());
if(model_file.is_open() == false)
{
log_error() << "Could not open the linear SVM model file " << filename << std::endl;
throw std::runtime_error("Failed to open the linear SVM model file");
}
parse_libsvm_model_file(model_file);
return;
}
/*
void NeuralClassifier::InitClassifier(){
// creating the mlp
ConnectedMachine mlp[10];
Linear *c1[10];
Tanh *c2[10];
Linear *c3[10];
Tanh *c4[10];
for(int count = 5; count <= 5; count++)
{
c1[count]= new(allocator) Linear(num_of_features,n_hu);
c2[count]= new(allocator) Tanh(n_hu);
c3[count]= new(allocator) Linear(n_hu,1);
c4[count]= new(allocator) Tanh(1);
mlp[count].addFCL(c1[count]);
mlp[count].addFCL(c2[count]);
mlp[count].addFCL(c3[count]);
mlp[count].addFCL(c4[count]);
mlp[count].build();
mlp[count].setPartialBackprop();
//mlp[count].loadXFile(&model_file);
}
}*/
void NeuralClassifier::TEST_DATA_BITS(char* test_file, char* model, int n_hu)
{
Allocator *allocator = new Allocator;
int num_of_features = 9;
int my_digit = 5;
cout<<"Testing.......... ."<<endl;
DiskXFile model_file(model, "r");
// creating the mlp
ConnectedMachine mlp[10];
Linear *c1[10];
Tanh *c2[10];
Linear *c3[10];
Tanh *c4[10];
for(int count = my_digit; count <= my_digit; count++)
{
c1[count]= new(allocator) Linear(num_of_features,n_hu);
c2[count]= new(allocator) Tanh(n_hu);
c3[count]= new(allocator) Linear(n_hu,1);
c4[count]= new(allocator) Tanh(1);
mlp[count].addFCL(c1[count]);
mlp[count].addFCL(c2[count]);
mlp[count].addFCL(c3[count]);
mlp[count].addFCL(c4[count]);
mlp[count].build();
mlp[count].setPartialBackprop();
mlp[count].loadXFile(&model_file);
}
cout<<"Reading data .......... ."<<endl;
DataSet *test_data =new(allocator) MatDataSet(test_file,num_of_features,1,false,-1,false);
/*Local variables & Initialization*/
float accuracy(0);
float positivesamples(0);
float n_iter=10;
Timer timer;
accuracy=0;
ofstream fout("test_stats.txt");
cout<<"Forwarding samples .......... ."<<endl;
long fivesamples=0;
/*Testing...*/
for(int t = 0; t < test_data->n_examples; t++)
{
if((t%100) == 0) cout<<t<<endl;
test_data->setExample(t);
int d = test_data->targets->frames[0][0];//actual value of sample
/*Get the score from the corresponding NN*/
mlp[my_digit].forward(test_data->inputs);
real score = mlp[my_digit].outputs->frames[0][0];
if (d==my_digit)
fivesamples++;
if(score > 0)
{
positivesamples++;
if(d == my_digit)
{
accuracy++;
}
else
{
char filename[50];
int writer, pass, digit;
writer = t/100;// the t 600 * 10 * 10 = writer 5670 /100 = 56
pass = (t-(writer*100))/10;
digit = (t-(writer*100))%10;
sprintf(filename, "writer%03u_pass%02u_digit%u.bmp", 601 + writer, 1 + pass, digit);
fout<<"### "<<t<<" "<<filename<<"; score = "<<score<<endl;
}
}
else if(d == my_digit)
{
char filename[50];
int writer, pass, digit;
writer = t/100;
pass = (t-(writer*100))/10;
digit = (t-(writer*100))%10;
sprintf(filename, "writer%03u_pass%02u_digit%u.bmp", 601 + writer, 1 + pass, digit);
fout<<"--- "<<t<<" "<<filename<<"; score = ["<<score<<"]"<<endl;
}
}
fout<<endl<<endl;
double samplerate=100.0/(double)fivesamples;
fout<<" total number of samples "<<test_data->n_examples<<"Count of digit 5 samples "<<fivesamples<<endl;
fout<<"Samples with result > 0: "<<positivesamples<<endl;
fout<<"Samples with the corresponding digit and result > 0: "<<accuracy<<"sample and "<< accuracy*samplerate<<" % "<<endl;
delete(allocator);
}
