abstractrandomforest.cpp

   1 //
   2 //  abstractrandomforest.cpp
   3 //  Mothur
   4 //
   5 //  Created by Sarah Westcott on 10/1/12.
   6 //  Copyright (c) 2012 Schloss Lab. All rights reserved.
   7 //
   8
   9 #include "abstractrandomforest.hpp"
  10
  11 /***********************************************************************/
  12 AbstractRandomForest::AbstractRandomForest(const std::vector < std::vector<int> > dataSet,
  13                      const int numDecisionTrees,
  14                      const string treeSplitCriterion = "informationGain")
  15 : dataSet(dataSet),
  16 numDecisionTrees(numDecisionTrees),
  17 numSamples((int)dataSet.size()),
  18 numFeatures((int)(dataSet[0].size() - 1)),
  19 globalDiscardedFeatureIndices(getGlobalDiscardedFeatureIndices()),
  20 globalVariableImportanceList(numFeatures, 0),
  21 treeSplitCriterion(treeSplitCriterion) {
  22     m = MothurOut::getInstance();
  23     // TODO: double check if the implemenatation of 'globalOutOfBagEstimates' is correct
  24 }
  25
  26 /***********************************************************************/
  27
  28 vector<int> AbstractRandomForest::getGlobalDiscardedFeatureIndices() {
  29     try {
  30         vector<int> globalDiscardedFeatureIndices;
  31
  32         // calculate feature vectors
  33         vector< vector<int> > featureVectors(numFeatures, vector<int>(numSamples, 0));
  34         for (int i = 0; i < numSamples; i++) {
  35             if (m->control_pressed) { return globalDiscardedFeatureIndices; }
  36             for (int j = 0; j < numFeatures; j++) { featureVectors[j][i] = dataSet[i][j]; }
  37         }
  38
  39         for (int i = 0; i < featureVectors.size(); i++) {
  40             if (m->control_pressed) { return globalDiscardedFeatureIndices; }
  41             double standardDeviation = m->getStandardDeviation(featureVectors[i]);
  42             if (standardDeviation <= 0){ globalDiscardedFeatureIndices.push_back(i); }
  43         }
  44
  45         if (m->debug) {
  46             m->mothurOut("number of global discarded features:  " + toString(globalDiscardedFeatureIndices.size())+ "\n");
  47             m->mothurOut("total features: " + toString(featureVectors.size())+ "\n");
  48         }
  49
  50         return globalDiscardedFeatureIndices;
  51     }
  52         catch(exception& e) {
  53                 m->errorOut(e, "AbstractRandomForest", "getGlobalDiscardedFeatureIndices");
  54                 exit(1);
  55         }
  56 }
  57
  58 /***********************************************************************/