Skip to content

Commit

Permalink
debugging, wip
Browse files Browse the repository at this point in the history
  • Loading branch information
Zeynep committed Aug 6, 2024
1 parent 6f39e2f commit 0661988
Showing 1 changed file with 15 additions and 16 deletions.
31 changes: 15 additions & 16 deletions ...eminf/clustering/art2a/clustering/euclideanClustering/Art2aEuclideanDoubleClustering.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -383,7 +383,6 @@ public IArt2aEuclideanClusteringResult getClusterResult(boolean anIsClusteringRe
tmpClusterMatrixRow[tmpCurrentVectorComponentsInClusterMatrixIndex] = tmpInitialClusterVectorWeightValue;
tmpClusterMatrixRowOld[tmpCurrentVectorComponentsInClusterMatrixIndex] = tmpInitialClusterVectorWeightValue;
}
System.out.println("Cluster matrix after initialization: " + Arrays.deepToString(this.clusterMatrix));
}
//</editor-fold>
//<editor-fold desc="Clustering results in text files set up." defaultstate="collapsed">
Expand Down Expand Up @@ -416,6 +415,7 @@ public IArt2aEuclideanClusteringResult getClusterResult(boolean anIsClusteringRe
tmpIsNullVector = false;
}
}
System.out.println("test input vector: " + Arrays.toString(tmpInputVector));
if(anIsClusteringResultExported) {
this.clusteringProcess.add(String.format("Input: %d / Vector %d", tmpCurrentInput,
tmpSampleVectorIndicesInRandomOrder[tmpCurrentInput]));
Expand All @@ -438,7 +438,6 @@ public IArt2aEuclideanClusteringResult getClusterResult(boolean anIsClusteringRe
tmpInputVector[tmpManipulateComponentsIndex] = 0.0;
}
}
LOGGER.info("Input vector after contrast enhancement: ");
//</editor-fold>
//<editor-fold desc="First pass, no clusters available, so the first cluster is created." defaultstate="collapsed">
if(tmpNumberOfDetectedClusters == 0) {
Expand All @@ -451,49 +450,46 @@ public IArt2aEuclideanClusteringResult getClusterResult(boolean anIsClusteringRe
this.clusteringProcess.add(String.format("Number of detected clusters: %d",tmpNumberOfDetectedClusters));
System.out.println("Cluster number: 0");
}
System.out.println("first cluster number: " + tmpNumberOfDetectedClusters);
}
//</editor-fold>
else {
tmpWinnerClusterIndex = tmpNumberOfDetectedClusters;
boolean tmpIsMatchingClusterAvailable = true;
//<editor-fold desc="Cluster number is greater than or equal to 1, so a rho winner is determined as shown in the following steps." defaultstate="collapsed">
//sum of scaling factor and non normalized components!
tmpRho = 0.0;
double tmpDifferenceOfScalingFactorAndInput = 0.0;
for (int tmpInputVectorComponent = 0; tmpInputVectorComponent < this.numberOfComponents; tmpInputVectorComponent++ ){
for (int tmpInputVectorComponent = 0; tmpInputVectorComponent < this.numberOfComponents; tmpInputVectorComponent++) {
tmpDifferenceOfScalingFactorAndInput = this.scalingFactor - tmpInputVector[tmpInputVectorComponent];
tmpRho += tmpDifferenceOfScalingFactorAndInput * tmpDifferenceOfScalingFactorAndInput;
System.out.println("Debug: Inside loop, tmpRho = " + tmpRho);
}
tmpWinnerClusterIndex = tmpNumberOfDetectedClusters;
boolean tmpIsMatchingClusterAvailable = true;
//tmpWinnerClusterIndex = tmpNumberOfDetectedClusters;
//boolean tmpIsMatchingClusterAvailable = true;
//<editor-fold desc="Calculate first rho value."
//tmpRho += tmpDifferenceOfScalingFactorAndInput * tmpDifferenceOfScalingFactorAndInput;
System.out.println("Calculated tmpRho: " + tmpRho);
//</editor-fold>
//<editor-fold desc="Calculation of the 2nd rho value and comparison of the two rho values to determine the rho winner."
for(int tmpCurrentClusterMatrixRow = 0; tmpCurrentClusterMatrixRow < tmpNumberOfDetectedClusters;
tmpCurrentClusterMatrixRow++) {
for (int tmpCurrentClusterMatrixRow = 0; tmpCurrentClusterMatrixRow < tmpNumberOfDetectedClusters;
tmpCurrentClusterMatrixRow++) {
double[] tmpRow;
double tmpRhoForExistingClustersSquared = 0.0;
double tmpDistance = 0.0;
tmpRow = this.clusterMatrix[tmpCurrentClusterMatrixRow];
for(int tmpElementsInRowIndex = 0; tmpElementsInRowIndex < this.numberOfComponents; tmpElementsInRowIndex++) {
for (int tmpElementsInRowIndex = 0; tmpElementsInRowIndex < this.numberOfComponents; tmpElementsInRowIndex++) {
double tmpRhoForExistingClusters = tmpInputVector[tmpElementsInRowIndex] - tmpRow[tmpElementsInRowIndex];
double tmpDifferenceSquared = Math.pow(tmpRhoForExistingClusters, 2);
tmpRhoForExistingClustersSquared += tmpDifferenceSquared;
//tmpRhoForExistingClustersSquared += tmpRhoForExistingClusters * tmpRhoForExistingClusters;
//tmpDistance += Math.pow(tmpInputVector[tmpElementsInRowIndex] - tmpRow[tmpElementsInRowIndex], 2);
System.out.println("index: " + tmpElementsInRowIndex + ", tmpRow[tmpElementsInRowIndex]: " + tmpRow[tmpElementsInRowIndex] + ", tmpInputVector[tmpElementsInRowIndex]: " +tmpInputVector[tmpElementsInRowIndex] + ": tmpRhoForExistingClusters = " + tmpRhoForExistingClusters +
", tmpDifferenceSquared = " + tmpDifferenceSquared + ", tmpRhoForExistingClustersSquared = " +tmpRhoForExistingClustersSquared);
}
System.out.println("Calculated tmpRhoForExistingClustersSquared for cluster " + tmpCurrentClusterMatrixRow + ": " + tmpRhoForExistingClustersSquared);
if(tmpRhoForExistingClustersSquared > tmpRho) {
if (tmpRhoForExistingClustersSquared > tmpRho) {
tmpRho = tmpRhoForExistingClustersSquared;
tmpWinnerClusterIndex = tmpCurrentClusterMatrixRow;
tmpIsMatchingClusterAvailable = false;
System.out.println("Debug: Updated tmpRho = " + tmpRho + ", tmpWinnerClusterIndex = " + tmpWinnerClusterIndex);
System.out.println("right if loop?: ");
}
LOGGER.info("Calculated tmpRho: " + tmpRho);
LOGGER.info("Calculated tmpRhoForExistingClustersSquared for cluster " + tmpCurrentClusterMatrixRow + ": " + tmpRhoForExistingClustersSquared);
}
//</editor-fold>
//<editor-fold desc="Deciding whether the input fits into an existing cluster or whether a new cluster must be formed."
Expand All @@ -503,6 +499,9 @@ public IArt2aEuclideanClusteringResult getClusterResult(boolean anIsClusteringRe
tmpNumberOfDetectedClusters++;
tmpClusterOccupation[tmpSampleVectorIndicesInRandomOrder[tmpCurrentInput]] =
tmpNumberOfDetectedClusters - 1;
System.out.println("cluster matrix size. " + this.clusterMatrix.length);
System.out.println(" number of detected clusters: " + Arrays.toString(this.clusterMatrix[4]));
System.out.println( " number of new detected clusters: " + tmpNumberOfDetectedClusters);
this.clusterMatrix[tmpNumberOfDetectedClusters - 1] = tmpInputVector;
if(anIsClusteringResultExported) {
this.clusteringProcess.add(String.format("Cluster number: %d",(tmpNumberOfDetectedClusters - 1)));
Expand Down

0 comments on commit 0661988

Please sign in to comment.