Hello,
I have been trying to do SVM classification in C++ and I used DAAL Library but unfortunately, I am getting wrong predictions.
Even when I test the algorithm in the same data that I trained the algorithm on I get, wrong predictions svm_two_class_train_dense.csv and svm_two_class_test_dense.csv have the same values but in a different order.
The code:
#include "daal.h"
#include "service.h"
using namespace daal;
using namespace daal::algorithms;
using namespace daal::data_management;
string trainDatasetFileName = "svm_two_class_train_dense.csv";
string testDatasetFileName = "svm_two_class_test_dense.csv";
const size_t nFeatures = 12;
/* Parameters for the SVM kernel function */
kernel_function::KernelIfacePtr kernel(new kernel_function::linear::Batch<>());
/* Model object for the SVM algorithm */
svm::training::ResultPtr trainingResult;
classifier::prediction::ResultPtr predictionResult;
NumericTablePtr testGroundTruth;
void trainModel();
void testModel();
void printResults();
int main(int argc, char** argv) {
string imageName;
double* p;
double test[F];
/*
*Write haralik_features results into CSV files
*/
//WriteToCSV();
trainModel();
testModel();
printResults();
return 0;
}
void trainModel()
{
/* Initialize FileDataSource<CSVFeatureManager> to retrieve the input data from a .csv file */
FileDataSource<CSVFeatureManager> trainDataSource(trainDatasetFileName, DataSource::notAllocateNumericTable, DataSource::doDictionaryFromContext);
/* Create Numeric Tables for training data and labels */
NumericTablePtr trainData = HomogenNumericTable<>::create(nFeatures, 0, NumericTable::doNotAllocate);
NumericTablePtr trainGroundTruth = HomogenNumericTable<>::create(1, 0, NumericTable::doNotAllocate);
NumericTablePtr mergedData = MergedNumericTable::create(trainData, trainGroundTruth);
/* Retrieve the data from the input file */
trainDataSource.loadDataBlock(mergedData.get());
/* Create an algorithm object to train the SVM model */
svm::training::Batch<float, svm::training::boser> algorithm;
algorithm.parameter.kernel = kernel;
algorithm.parameter.cacheSize = 40000000;
/* Pass a training data set and dependent values to the algorithm */
algorithm.input.set(classifier::training::data, trainData);
algorithm.input.set(classifier::training::labels, trainGroundTruth);
/* Build the SVM model */
algorithm.compute();
/* Retrieve the algorithm results */
trainingResult = algorithm.getResult();
}
void testModel()
{
/* Initialize FileDataSource<CSVFeatureManager> to retrieve the test data from a .csv file */
FileDataSource<CSVFeatureManager> testDataSource(testDatasetFileName, DataSource::notAllocateNumericTable, DataSource::doDictionaryFromContext);
/* Create Numeric Tables for testing data and labels */
NumericTablePtr testData = HomogenNumericTable<>::create(nFeatures, 0, NumericTable::doNotAllocate);
testGroundTruth = HomogenNumericTable<>::create(1, 0, NumericTable::doNotAllocate);
NumericTablePtr mergedData = MergedNumericTable::create(testData, testGroundTruth);
/* Retrieve the data from input file */
testDataSource.loadDataBlock(mergedData.get());
/* Create an algorithm object to predict SVM values */
svm::prediction::Batch<> algorithm;
algorithm.parameter.kernel = kernel;
/* Pass a testing data set and the trained model to the algorithm */
algorithm.input.set(classifier::prediction::data, testData);
algorithm.input.set(classifier::prediction::model, trainingResult->get(classifier::training::model));
/* Predict SVM values */
algorithm.compute();
/* Retrieve the algorithm results */
predictionResult = algorithm.getResult();
}
void printResults()
{
printNumericTables<int, float>(testGroundTruth, predictionResult->get(classifier::prediction::prediction), "Ground truth\t",
"Classification results", "SVM classification results (first 11 observations):", 11);
}
The output:
Any help?
Thank you
链接已复制
