Cluster prior set according to graph size

Author: smm8ec

src/main/java/edu/virginia/uvacluster/internal/MyControlPanel.java

@@ -488,6 +488,7 @@ public void actionPerformed(ActionEvent e) {
 										numSeeds.setText(Integer.toString(numNodes / 8));
 									}
 								}
+								setClusterPrior();
 					    	}
 					    }
 					  }
@@ -1151,12 +1152,13 @@ private void evaluateButtonPressed(File evaluateFile) throws IOException {
 	}
 
 	private Double minScoreThreshold() {
-		if (proteinGraph.getSelectedItem().toString().equals(" - Select Network - ")) {
+		String proteinGraphName = proteinGraph.getSelectedItem().toString();
+		if (proteinGraphName.equals(" - Select Network - ")) {
 			JOptionPane.showMessageDialog(this, "Please select a protein graph under 'Search'");
 		} else {
 			CyNetwork nw = null;
 			for (CyNetwork network: CyActivator.networkManager.getNetworkSet()) {
-				if (network.getRow(network).get(CyNetwork.NAME, String.class).equals(proteinGraph.getSelectedItem().toString())) {
+				if (network.getRow(network).get(CyNetwork.NAME, String.class).equals(proteinGraphName)) {
 					nw = network; 
 					break;
 				}
@@ -1259,6 +1261,26 @@ private InputTask createInputTask() {
 		return inputTask;
 	}
 
+	private int getProteinGraphSize(String proteinGraphName) {
+		CyNetwork nw = null;
+		for (CyNetwork network: CyActivator.networkManager.getNetworkSet()) {
+			if (network.getRow(network).get(CyNetwork.NAME, String.class).equals(proteinGraphName)) {
+				nw = network; 
+				break;
+			}
+		}	
+		return nw.getNodeCount();
+	}
+	
+	private void setClusterPrior() {
+		int graphSize = getProteinGraphSize(proteinGraph.getSelectedItem().toString());
+		if (graphSize <= 1000) {
+			clusterPrior.setText("0.5");
+		} else {
+			clusterPrior.setText("0.0001");
+		}
+	}
+	
 	private Integer validateInput() {
 
 		if (learningScoreOption.isSelected() && !trainDefaultModel.isSelected() && !trainCustomModel.isSelected()) {

src/main/java/edu/virginia/uvacluster/internal/SupervisedModel.java

@@ -284,73 +284,17 @@ private CySubNetwork genNegativeExample(int size, List<CySubNetwork> positiveExa
 	 * Convenience method for training networks on positive *and* negative examples
 	 */
 	public void train(List<CySubNetwork> positiveExamples, List<CySubNetwork> negativeExamples) {
-		System.out.println("Entered TRAIN");
-		
-		for (FeatureSet feature : features) {
-			for (String desc : feature.getDescriptions()) {
-				System.out.println(desc);
-			}
-		}
-		
 		List<Cluster> posExamples = new ArrayList<Cluster>(), negExamples = new ArrayList<Cluster>();
 		for(CySubNetwork pos: positiveExamples) {posExamples.add(new Cluster(features, pos)); }
 		for(CySubNetwork neg: negativeExamples) {negExamples.add(new Cluster(features, neg));}
 
 		System.out.println("Lists of pos and neg training examples created");
-		
-//		System.out.println("TRAINING BINS ON POSITIVE COMPLEXES:");
+
 		posBayesGraph.trainBins(posExamples);
-//		for (FeatureSet feature : features) {
-//			List<String> statNames = feature.getDescriptions();
-//			Map<String, Statistic> statMap = feature.getStatisticMap();
-//			for (String statName : statNames) {
-//				Statistic stat = statMap.get(statName);
-//				Double min = stat.getRange().getMin();
-//				Double max = stat.getRange().getMax();
-//				System.out.println(statName + "\n\tMin: " + min + "\n\tMax: " + max);
-//			}
-//		}
-	
-//		System.out.println("TRAINING BINS ON NEGATIVE COMPLEXES:");
 		negBayesGraph.trainBins(negExamples);
-//		for (FeatureSet feature : features) {
-//			List<String> statNames = feature.getDescriptions();
-//			Map<String, Statistic> statMap = feature.getStatisticMap();
-//			for (String statName : statNames) {
-//				Statistic stat = statMap.get(statName);
-//				Double min = stat.getRange().getMin();
-//				Double max = stat.getRange().getMax();
-//				System.out.println(statName + "\n\tMin: " + min + "\n\tMax: " + max);
-//			}
-//		}
-		
-//		System.out.println("TRAINING PROBABILITY  - POSITIVE:");
-		posBayesGraph.trainOn(posExamples);
-		// Min/max stable
-//		for (FeatureSet feature : features) {
-//			List<String> statNames = feature.getDescriptions();
-//			Map<String, Statistic> statMap = feature.getStatisticMap();
-//			for (String statName : statNames) {
-//				Statistic stat = statMap.get(statName);
-//				Double min = stat.getRange().getMin();
-//				Double max = stat.getRange().getMax();
-//				System.out.println(statName + "\n\tMin: " + min + "\n\tMax: " + max);
-//			}
-//		}
-		
-//		System.out.println("TRAINING PROBABILITY - NEGATIVE:");		
+		posBayesGraph.trainOn(posExamples);	
 		negBayesGraph.trainOn(negExamples);
-		// Min/max stable
-//		for (FeatureSet feature : features) {
-//			List<String> statNames = feature.getDescriptions();
-//			Map<String, Statistic> statMap = feature.getStatisticMap();
-//			for (String statName : statNames) {
-//				Statistic stat = statMap.get(statName);
-//				Double min = stat.getRange().getMin();
-//				Double max = stat.getRange().getMax();
-//				System.out.println(statName + "\n\tMin: " + min + "\n\tMax: " + max);
-//			}
-//		}
+
 		System.out.println("Model has finished training on " + negativeExamples.size() +  " negative Examples.");
 	}
 

src/main/java/edu/virginia/uvacluster/internal/feature/FeatureUtil.java

@@ -48,8 +48,8 @@ public static List<FeatureSet> parse(Set<String> featureKeys) {
 			m.matches();
 			numBins = Integer.parseInt(m.group(2));
 
-			// Check if Mean : weight -- to use alternative mean calculation
-			// over all possible edges
+			/* Check if Mean : weight -- to use alternative mean calculation
+			 over all possible edges */
 			if (weightFeaturePattern.matcher(featureName).matches() && statName.equals("Mean"))
 				statName = "mean possible";