ENH: add option to choose between margin and confidence in RF classification

Modules/Learning/Supervised/include/otbRandomForestsMachineLearningModel.h

@@ -121,6 +121,9 @@ public:
   itkGetMacro(TerminationCriteria, int);
   itkSetMacro(TerminationCriteria, int);
 
+  itkGetMacro(CoputeMargin, bool);
+  itkSetMacro(ComputeMargin, bool);
+
   /** Returns a matrix containing variable importance */
   VariableImportanceMatrixType GetVariableImportance();
   
@@ -206,6 +209,10 @@ private:
   float m_ForestAccuracy;
   /** The type of the termination criteria */
   int m_TerminationCriteria;
+  /** Wether to compute margin (difference in probability between the
+   * 2 most voted classes) instead of confidence (probability of the most
+   * voted class) in prediction*/
+  bool m_ComputeMargin;
 };
 } // end namespace otb
 

Modules/Learning/Supervised/include/otbRandomForestsMachineLearningModel.txx

@@ -39,7 +39,8 @@ RandomForestsMachineLearningModel<TInputValue,TOutputValue>
   m_MaxNumberOfVariables(0),
   m_MaxNumberOfTrees(100),
   m_ForestAccuracy(0.01),
-  m_TerminationCriteria(CV_TERMCRIT_ITER | CV_TERMCRIT_EPS)
+  m_TerminationCriteria(CV_TERMCRIT_ITER | CV_TERMCRIT_EPS),
+  m_ComputeMargin(false)
 {
   this->m_ConfidenceIndex = true;
   this->m_IsRegressionSupported = true;
@@ -91,7 +92,7 @@ RandomForestsMachineLearningModel<TInputValue,TOutputValue>
                                  m_MaxNumberOfTrees,            // max number of trees in the forest
                                  m_ForestAccuracy,              // forest accuracy
                                  m_TerminationCriteria          // termination criteria
-                                 );
+    );
 
   cv::Mat var_type = cv::Mat(this->GetInputListSample()->GetMeasurementVectorSize() + 1, 1, CV_8U );
   var_type.setTo(cv::Scalar(CV_VAR_NUMERICAL) ); // all inputs are numerical
@@ -125,7 +126,10 @@ RandomForestsMachineLearningModel<TInputValue,TOutputValue>
 
   if (quality != NULL)
     {
-    (*quality) = m_RFModel->predict_confidence(sample);
+    if(m_ComputeMargin)
+      (*quality) = m_RFModel->predict_margin(sample);
+    else
+      (*quality) = m_RFModel->predict_confidence(sample);
     }
 
   return target[0];
@@ -176,12 +180,12 @@ RandomForestsMachineLearningModel<TInputValue,TOutputValue>
     //if (line.find(m_RFModel->getName()) != std::string::npos)
     if (line.find(CV_TYPE_NAME_ML_RTREES) != std::string::npos)
       {
-         //std::cout<<"Reading a "<<CV_TYPE_NAME_ML_RTREES<<" model"<<std::endl;
-         return true;
+      //std::cout<<"Reading a "<<CV_TYPE_NAME_ML_RTREES<<" model"<<std::endl;
+      return true;
       }
-   }
-   ifs.close();
-   return false;
+    }
+  ifs.close();
+  return false;
 }
 
 template <class TInputValue, class TOutputValue>