X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=randomforest.cpp;h=acf87dfebcd022d37cf6974f8331c7ad940a017a;hp=1fcd23df79f8b925a0ceede3245cda94b7f2a884;hb=499f4ac6e321f9f03d4c3aa25c3b6880892c8b83;hpb=5072e9b05905ef385226a634b98189b4f9781489

diff --git a/randomforest.cpp b/randomforest.cpp
index 1fcd23d..acf87df 100644
--- a/randomforest.cpp
+++ b/randomforest.cpp
@@ -104,6 +104,38 @@ int RandomForest::printConfusionMatrix(map<int, string> intToTreatmentMap) {
 	}
 }
 
+/***********************************************************************/
+
+int RandomForest::getMissclassifications(string filename, map<int, string> intToTreatmentMap, vector<string> names) {
+    try {
+        ofstream out;
+        m->openOutputFile(filename, out);
+        out <<"Sample\tRF classification\tActual classification\n";
+        for (map<int, vector<int> >::iterator it = globalOutOfBagEstimates.begin(); it != globalOutOfBagEstimates.end(); it++) {
+            
+            if (m->control_pressed) { return 0; }
+            
+            int indexOfSample = it->first;
+            vector<int> predictedOutComes = it->second;
+            vector<int>::iterator maxPredictedOutComeIterator = max_element(predictedOutComes.begin(), predictedOutComes.end());
+            int majorityVotedOutcome = (int)(maxPredictedOutComeIterator - predictedOutComes.begin());
+            int realOutcome = dataSet[indexOfSample][numFeatures];
+                                   
+            if (majorityVotedOutcome != realOutcome) {             
+                out << names[indexOfSample] << "\t" << intToTreatmentMap[majorityVotedOutcome] << "\t" << intToTreatmentMap[realOutcome] << endl;
+                                
+            }
+        }
+        
+        out.close();    
+        return 0;
+    }
+	catch(exception& e) {
+		m->errorOut(e, "RandomForest", "getMissclassifications");
+		exit(1);
+	} 
+}
+
 /***********************************************************************/
 int RandomForest::calcForrestVariableImportance(string filename) {
     try {
@@ -148,7 +180,7 @@ int RandomForest::calcForrestVariableImportance(string filename) {
         m->openOutputFile(filename, out);
         out <<"OTU\tMean decrease accuracy\n";
         for (int i = 0; i < globalVariableRanks.size(); i++) {
-            out << m->currentBinLabels[(int)globalVariableRanks[i].first] << '\t' << globalVariableImportanceList[globalVariableRanks[i].first] << endl;
+            out << m->currentSharedBinLabels[(int)globalVariableRanks[i].first] << '\t' << globalVariableImportanceList[globalVariableRanks[i].first] << endl;
         }
         out.close();
         return 0;