RegressionData.h

Go to the documentation of this file.

#ifndef GRT_REGRESSION_DATA_HEADER
#define GRT_REGRESSION_DATA_HEADER

#include "../Util/GRTTypedefs.h"
 #include "VectorFloat.h"
#include "RegressionSample.h"

GRT_BEGIN_NAMESPACE

class GRT_API RegressionData{
public:
    
    RegressionData(const UINT numInputDimensions=0,const UINT numTargetDimensions=0,const std::string datasetName="NOT_SET",const std::string infoText="");
    
 RegressionData(const RegressionData &rhs);
    
 ~RegressionData();

 RegressionData& operator=(const RegressionData &rhs);
 
 inline RegressionSample& operator[](const UINT &i){
  return data[i];
 }

    const inline RegressionSample& operator[](const UINT &i) const{
        return data[i];
    }
    
 void clear();

    bool setInputAndTargetDimensions(const UINT numInputDimensions,const UINT numTargetDimensions);
    
    bool setDatasetName(const std::string &datasetName);
    
    bool setInfoText(const std::string &infoText);
    
 bool addSample(const VectorFloat &inputVector,const VectorFloat &targetVector);
    
 bool removeLastSample();
    
    bool reserve(const UINT N);
    
    bool setExternalRanges(const Vector< MinMax > &externalInputRanges, const Vector< MinMax > &externalTargetRanges, const bool useExternalRanges);
    
    bool enableExternalRangeScaling(const bool useExternalRanges);
    
    bool scale(const Float minTarget,const Float maxTarget);
    
    bool scale(const Vector< MinMax > &inputVectorRanges,const Vector< MinMax > &targetVectorRanges,const Float minTarget,const Float maxTarget);
    
    bool save(const std::string &filename) const;
    
    bool load(const std::string &filename);
    
    bool saveDatasetToFile(const std::string &filename) const;
 
    bool loadDatasetFromFile(const std::string &filename);
    
    bool saveDatasetToCSVFile(const std::string &filename) const;
 
    bool loadDatasetFromCSVFile(const std::string &filename,const UINT numInputDimensions,const UINT numTargetDimensions);
    
    bool printStats() const;
    
    bool merge(const RegressionData &regressionData);
    
    GRT_DEPRECATED_MSG( "partition(...) is deprecated, use split(...) instead", RegressionData partition(const UINT trainingSizePercentage) );

    RegressionData split(const UINT trainingSizePercentage);
    
    bool spiltDataIntoKFolds(const UINT K);
    
    RegressionData getTrainingFoldData(const UINT foldIndex) const;
    
    RegressionData getTestFoldData(const UINT foldIndex) const;

    UINT removeDuplicateSamples();
    
    std::string getDatasetName()const { return datasetName; }
    
    std::string getInfoText() const { return infoText; }

    std::string getStatsAsString() const;
    
    UINT inline getNumInputDimensions() const{ return numInputDimensions; }
    
    UINT inline getNumTargetDimensions() const{ return numTargetDimensions; }
 
    UINT inline getNumSamples() const{ return totalNumSamples; }
    
    Vector<MinMax> getInputRanges() const;
    
    Vector<MinMax> getTargetRanges() const;
    
    Vector< RegressionSample > getData() const{ return data; }

private:
    std::string datasetName;                                     
    std::string infoText;                                        
 UINT numInputDimensions;                                
 UINT numTargetDimensions;                               
 UINT totalNumSamples;                                   
    UINT kFoldValue;                                        
    bool crossValidationSetup;                              
    bool useExternalRanges;                                 
    Vector< MinMax > externalInputRanges;                   
    Vector< MinMax > externalTargetRanges;                  
 Vector< RegressionSample > data;                        
    Vector< Vector< UINT > >    crossValidationIndexs;      
    
    DebugLog debugLog;                                      
    ErrorLog errorLog;                                      
    WarningLog warningLog;                                  
};

GRT_END_NAMESPACE

#endif //GRT_REGRESSION_DATA_HEADER