performance improvement: filtering the id attribute before

classification

Author: piskachev

.classpath

@@ -3,8 +3,7 @@
 	<classpathentry kind="con" path="org.eclipse.jdt.launching.JRE_CONTAINER"/>
 	<classpathentry kind="src" path="src"/>
 	<classpathentry kind="lib" path="lib/json-simple-1.1.1.jar"/>
-	<classpathentry kind="lib" path="lib/soot-infoflow-trunk.jar"/>
-	<classpathentry kind="lib" path="lib/soot.jar"/>
 	<classpathentry kind="lib" path="lib/weka.jar"/>
+	<classpathentry kind="lib" path="lib/soot-infoflow-trunk.jar"/>
 	<classpathentry kind="output" path="bin"/>
 </classpath>

bin/de/fraunhofer/iem/mois/Learner.class

@@ -22,6 +22,7 @@
 import weka.classifiers.bayes.NaiveBayes;
 import weka.classifiers.functions.Logistic;
 import weka.classifiers.functions.SMO;
+import weka.classifiers.meta.FilteredClassifier;
 import weka.classifiers.rules.JRip;
 import weka.classifiers.rules.OneR;
 import weka.classifiers.trees.DecisionStump;
@@ -32,6 +33,7 @@
 import weka.core.Instances;
 import weka.core.Range;
 import weka.core.converters.ArffSaver;
+import weka.filters.unsupervised.attribute.Remove;
 
 /**
  * Finds possible sources and sinks in a given set of system methods using a
@@ -46,8 +48,8 @@ public class Learner {
   private long startAnalysisTime;
   private long analysisTime;
 
-  private final boolean CROSS_EVALUATE=true;
-  private final boolean CLASSIFY=false;
+  private final boolean CROSS_EVALUATE=false;
+  private final boolean CLASSIFY=true;
 
   private final int CROSS_EVALUATE_ITERATIONS=1;
 
@@ -61,6 +63,7 @@ public double classify(Set<Method> trainingSet, Set<Method> testSet,
       Map<Category, Set<IFeature>> features, Set<Category> categories,
       String outputFile, boolean cweMode) throws IOException {
 
+	  
 	  double fmeasure=0;
     startAnalysisTime = System.currentTimeMillis();
     Map<Category, Integer> counters = new HashMap<Category, Integer>();
@@ -186,32 +189,36 @@ public double classify(Set<Method> trainingSet, Set<Method> testSet,
     }
     //System.out.println("Done.");
 
+         
+     
+	     
     // Create classifier.
     try {
       // instances.randomize(new Random(1337));
-      Classifier classifier = null;
+    	FilteredClassifier classifier = new FilteredClassifier();
       // (IBK / kNN) vs. SMO vs. (J48 vs. JRIP) vs. NaiveBayes
       // MultiClassClassifier für ClassifierPerformanceEvaluator
       if (WEKA_LEARNER_ALL.equals("BayesNet"))
-        classifier = new BayesNet();
+        classifier.setClassifier(new BayesNet());
       else if (WEKA_LEARNER_ALL.equals("NaiveBayes"))
-        classifier = new NaiveBayes();
+        classifier.setClassifier(new NaiveBayes());
       else if (WEKA_LEARNER_ALL.equals("J48"))
-        classifier = new J48();
+        classifier.setClassifier(new J48());
       else if (WEKA_LEARNER_ALL.equals("SMO"))
-        classifier = new SMO();
+        classifier.setClassifier(new SMO());
       else if (WEKA_LEARNER_ALL.equals("JRip"))
-        classifier = new JRip();
+        classifier.setClassifier(new JRip());
       else if (WEKA_LEARNER_ALL.equals("DecisionStump"))
-          classifier = new DecisionStump();
+          classifier.setClassifier(new DecisionStump());
       else if (WEKA_LEARNER_ALL.equals("OneR"))
-          classifier = new OneR();
+          classifier.setClassifier(new OneR());
       else if (WEKA_LEARNER_ALL.equals("Logistic"))
-          classifier = new Logistic();
+          classifier.setClassifier(new Logistic());
       else
         throw new Exception("Wrong WEKA learner!");
      // System.out.println("Classifier created: " + WEKA_LEARNER_ALL);
 
+      
       // Save arff data.
       ArffSaver saver = new ArffSaver();
       saver.setInstances(trainInstances);
@@ -259,15 +266,23 @@ sb, new Range(attributes.indexOf(idAttr) + 1
       // Classification.
 
       if(CLASSIFY) {
+    	  
+    	  Remove rm = new Remove();      
+      	rm.setAttributeIndicesArray(new int[]{idAttr.index()}); 
+      	classifier.setFilter(rm);   
+    	  
     	  //System.out.println("Classification starting.");
           classifier.buildClassifier(trainInstances);
           if (WEKA_LEARNER_ALL.equals("J48")) {
-            System.out.println(((J48) (classifier)).graph());
+            System.out.println(((classifier)).graph());
           }
           for (int instIdx = 0; instIdx < testInstances.numInstances(); instIdx++) {
             Instance inst = testInstances.instance(instIdx);
             assert inst.classIsMissing();
             Method meth = instanceMethods.get(inst.stringValue(idAttr));
+            
+            
+            
             double d = classifier.classifyInstance(inst);
             String cName = testInstances.classAttribute().value((int) d);
             boolean found = false;
@@ -297,7 +312,7 @@ sb, new Range(attributes.indexOf(idAttr) + 1
 
     Runtime.getRuntime().gc();
     analysisTime = System.currentTimeMillis() - startAnalysisTime;
-    //System.out.println("Time to classify " + categories.toString() + ": "+ analysisTime + " ms");
+    System.out.println("Time to classify " + categories.toString() + ": "+ analysisTime + " ms");
 
     return fmeasure;
     // writer.writeRIFLSpecification(outputFile, methods);

bin/de/fraunhofer/iem/mois/Main.class

@@ -49,10 +49,10 @@ public static void main(String[] args) {
   private static final boolean runSources = true;
   private static final boolean runSinks = true;
   private static final boolean runSanitizers = true;
-  private static final boolean runAuthentications = true;
+  private static final boolean runAuthentications = false;
   private static final boolean runCwes = true;
 
-  private static final boolean runOAT = true; // run one at a time analysis
+  private static final boolean runOAT = false; // run one at a time analysis
 
   private void run(String[] args) throws IOException {
 	  int iterations = 0;
@@ -144,7 +144,7 @@ private void run(String[] args) throws IOException {
 		        iter++;
 		      }
 		    }
-		    System.out.println("***** F Measure is " + averageF/iter);
+		    //System.out.println("***** F Measure is " + averageF/iter);
 
 		    //System.out.println("***** Writing final results");
 		    Set<String> tmpFiles = Util.getFiles(args[3]);

bin/de/fraunhofer/iem/mois/data/Category.class

@@ -14,10 +14,9 @@ public enum Category {
             Constants.AUTHENTICATION_NOCHANGE, false), SANITIZER(Constants.SANITIZER,
             false), NONE(Constants.NONE, false),
 
-  CWE089("CWE089", true), CWE306("CWE306", true), CWE078("CWE078",
+  CWE089("CWE089", true),CWE306("CWE306", true), CWE078("CWE078",
       true), CWE862("CWE862", true), CWE863("CWE863", 
           true), CWE601("CWE601", true), CWETEST("CWEtest", true), CWE079("CWE079",true), CWE_NONE("none", true);
-
   private final String id;
   private final boolean cwe;