Solved rare problem where repeated initial centroids could be created…

…, devoiding one of the after the 2nd iteration (and thus generating incorrect imputations)

src/keel/Algorithms/Preprocess/Missing_Values/kmeansImpute/gCenter.java

@@ -6,10 +6,10 @@
 	Copyright (C) 2004-2010
 	
 	F. Herrera ([email protected])
-    L. Sánchez ([email protected])
-    J. Alcalá-Fdez ([email protected])
-    S. García ([email protected])
-    A. Fernández ([email protected])
+    L. Sánchez ([email protected])
+    J. Alcalá-Fdez ([email protected])
+    S. García ([email protected])
+    A. Fernández ([email protected])
     J. Luengo ([email protected])
 
 	This program is free software: you can redistribute it and/or modify
@@ -29,7 +29,7 @@ J. Luengo ([email protected])
 
 /**
  * <p>
- * @author Written by Julián Luengo Martín 28/11/2006
+ * @author Written by Julián Luengo Martín 28/11/2006
  * @version 0.1
  * @since JDK 1.5
  * </p>
@@ -247,7 +247,7 @@ public void recalculateCenters(InstanceSet IS) {
         for (int a = 0; a < numCenters; a++) {
             for (int b = 0; b < nvariables; b++) {
             	nInst[a][b] = 0;
-                gravCenters[a][b] = "a";
+                gravCenters[a][b] = null;
                 modes[a][b] = new FreqList();
             }
         }
@@ -268,7 +268,7 @@ public void recalculateCenters(InstanceSet IS) {
                     if (tipo != Attribute.NOMINAL
                             && !i.getInputMissingValues(in)) {
                     	nInst[c][l]++;
-                    	if(gravCenters[c][l].compareTo("a") == 0)
+                    	if(gravCenters[c][l]==null)
                     		gravCenters[c][l] = new String("0");
                         tmp = new Double(gravCenters[c][l]).doubleValue();
                         tmp += i.getInputRealValues(in);
@@ -286,7 +286,7 @@ public void recalculateCenters(InstanceSet IS) {
                         if (tipo != Attribute.NOMINAL
                                 && !i.getOutputMissingValues(out)) {
                         	nInst[c][l]++;
-                        	if(gravCenters[c][l].compareTo("a") == 0)
+                        	if(gravCenters[c][l]==null)
                         		gravCenters[c][l] = new String("0");
                             tmp = new Double(gravCenters[c][l]).doubleValue();
                             tmp += i.getOutputRealValues(out);
@@ -326,7 +326,7 @@ public void recalculateCenters(InstanceSet IS) {
             tipo = at.getType();
             if (tipo != Attribute.NOMINAL) {
                 for (int a = 0; a < numCenters; a++) {
-                	if(gravCenters[a][b].compareTo("a") != 0){
+                	if(gravCenters[a][b]!=null){
                 		tmp = new Double(gravCenters[a][b]).doubleValue();
                 		tmp = tmp / nInst[a][b];
                 		gravCenters[a][b] = String.valueOf(tmp);

src/keel/Algorithms/Preprocess/Missing_Values/kmeansImpute/kmeansImpute.java

@@ -6,10 +6,10 @@
 	Copyright (C) 2004-2010
 	
 	F. Herrera ([email protected])
-    L. Sánchez ([email protected])
-    J. Alcalá-Fdez ([email protected])
-    S. García ([email protected])
-    A. Fernández ([email protected])
+    L. Sánchez ([email protected])
+    J. Alcalá-Fdez ([email protected])
+    S. García ([email protected])
+    A. Fernández ([email protected])
     J. Luengo ([email protected])
 
 	This program is free software: you can redistribute it and/or modify
@@ -29,7 +29,7 @@ J. Luengo ([email protected])
 
 /**
  * <p>
- * @author Written by Julián Luengo Martín 29/11/2006
+ * @author Written by Julián Luengo Martín 29/11/2006
  * @version 0.2
  * @since JDK 1.5
  * </p>
@@ -235,10 +235,8 @@ private double distance(Instance i,Instance j){
      */
     public void process(){
         //declarations
-        double []outputs;
-        double []outputs2;
-        Instance neighbor;
-        double dist,mean;
+    		ArrayList<Integer> initialCenters;
+	    	boolean repeated;
         int actual;
         Randomize rnd = new Randomize();
         Instance ex;
@@ -272,6 +270,7 @@ public void process(){
             //first, we choose k 'means' randomly from all
             //instances
             totalMissing = 0;
+            initialCenters = new ArrayList<Integer>();
             for(int i = 0;i < ndatos;i++){
                 Instance inst = IS.getInstance(i);
                 if(inst.existsAnyMissingValue())
@@ -285,8 +284,13 @@ public void process(){
                 do{
                     actual = (int) (ndatos*rnd.Rand());
                     ex = IS.getInstance(actual);
-                }while(ex.existsAnyMissingValue() && !allMissing);
-
+                    repeated = false;
+                    for(int i = 0; !ex.existsAnyMissingValue() && !repeated && i<initialCenters.size();i++){
+                    	if(initialCenters.get(i) == actual)
+                    		repeated = true;
+                    }
+                }while(ex.existsAnyMissingValue() && !allMissing || repeated);
+                initialCenters.add(actual);
                 kmeans.copyCenter(ex,numMeans);
             }