public void assignClusterClass(Cluster cluster) {
List<String> classString = cluster.getColumnAtString(cluster.getAttributes().size() - 1);
Set<String> set = new TreeSet<>();
for (String str : classString) {
set.add(str);
}
String name = null;
try {
name = set.toArray(new String[0])[0];
} catch (ArrayIndexOutOfBoundsException ae) {
TrainingLog.mainLogs.info("Array index out of bounds excepiton");
cluster.setClassName(Notations.DEFAULT_CLASS);
return;
}
int count = CommonLogics.numberOfInstance(classString, name);
Iterator<String> iter = set.iterator();
while (iter.hasNext()) {
String temp = iter.next();
int cct = CommonLogics.numberOfInstance(classString, temp);
if (cct > count) {
name = temp;
count = cct;
}
}
cluster.setClassName(name);
}
public List<String[]> findPoints(Cluster data, List<AttributeCluster> attrbs) {
List<String[]> list = new ArrayList<String[]>();
for (int i = 0; i < data.size(); i++) {
list.add(getPoints(attrbs, data.getRecords().get(i)));
}
return list;
}
public double findEucleideanDistance(String[] p1, Cluster cluster) {
double sum = 0.0;
for (int i = 0; i < cluster.size(); i++) {
sum += findEucleideanDistance(p1, cluster.getRecordPointsAt(i));
}
return sum;
}
/**
* Finds all the pairs of eucleidean distances between cluster 1 and cluster 2
*
* @param cluster 1
* @param cluster 2
* @return list of all distances
*/
public List<Double> findAllEucleideanDistances(Cluster cl1, Cluster cl2) {
List<Double> values = new ArrayList<Double>();
for (int i = 0; i < cl1.size(); i++) {
for (int j = 0; j < cl2.size(); j++) {
values.add(findEucleideanDistance(cl1.getRecordPointsAt(i), cl2.getRecordPointsAt(j)));
}
}
return values;
}
private void alterCentroid(Cluster cluster, List<AttributeCluster> attrb) {
if (cluster.size() > 0) {
List<Double> attr = new ArrayList<>();
for (int i = 0; i < attrb.size(); i++) {
if (attrb.get(i).getType() == TypeAttribute.continuous) {
List<Double> mean = cluster.getColumnAt(i);
attr.add(new CommonLogics().mean(mean));
}
}
cluster.setCentroid(new Centroid(attr));
}
}
/**
* Calculates SSE
*
* @param cluster
*/
public void calculateSSEForCluster(Cluster cluster) {
double SSE = 0.0;
for (int i = 0; i < cluster.size(); i++) {
double temp =
findEucleideanDistance(
cluster.getCentroid().getAttrList(),
getPoints(cluster.getAttributes(), cluster.getRecords().get(i)));
SSE += temp * temp;
}
cluster.setSSE(SSE);
}
public void recomputingCentroids(BisectCluster bisect, Cluster data) {
CluseterList list = new CluseterList();
list.addCluster(bisect.getC1());
list.addCluster(bisect.getC2());
for (int i = 0; i < list.size(); i++) {
alterCentroid(list.getClusterAt(i), data.getAttributes());
}
}
/**
* Add points from cluster to Bisect Cluster data model
*
* @param bisect
* @param data
*/
public void addClusterPoints(BisectCluster bisect, Cluster data) {
CluseterList list = new CluseterList();
list.addCluster(bisect.getC1());
list.addCluster(bisect.getC2());
list.clearsClusterPoints();
for (int i = 0; i < data.size(); i++) {
int temp =
findClosestClusterAndAddPoint(
list, getPoints(data.getAttributes(), data.getRecords().get(i)));
list.getClusterAt(temp).addPoints(data.getRecords().get(i));
}
calculateSSEForCluster(bisect.getC1());
// System.out.println(bisect.getC1().getSSE());
calculateSSEForCluster(bisect.getC2());
bisect.setTotalSSE(bisect.getC1().getSSE() + bisect.getC2().getSSE());
}
/**
* Bisects a cluster to give two cluster in a list
*
* @param cluster
* @return
* @throws CloneNotSupportedException
*/
public List<Cluster> bisectCluster(Cluster cluster) throws CloneNotSupportedException {
Cluster c1 =
new Cluster(new CommonLogics().generateRandomNumbers(1, cluster.numberOfPoints()), 0);
Cluster c2 =
new Cluster(new CommonLogics().generateRandomNumbers(1, cluster.numberOfPoints()), 0);
c1.setAttributes(cluster.getAttributes());
c2.setAttributes(cluster.getAttributes());
BisectClusterList blist = new BisectClusterList();
BisectCluster bisect = new BisectCluster(c1.clone(), c2.clone());
addClusterPoints(bisect, cluster);
blist.addBisectCluster(bisect);
for (int i = 0; i < ValueConstants.TRIALS; i++) {
Cluster cc1 = bisect.getC1().clone();
Cluster cc2 = bisect.getC2().clone();
BisectCluster bisectt = new BisectCluster(cc1, cc2);
recomputingCentroids(bisectt, cluster);
addClusterPoints(bisectt, cluster);
blist.addBisectCluster(bisectt);
// blist.getBisectList().get(i).setTotalSSE(bisectt.getC1().getSSE()+bisectt.getC2().getSSE());
}
int lowest = 0;
for (int i = 0; i < blist.size(); i++) {
if (blist.SSEat(i) < blist.SSEat(lowest)) {
lowest = i;
}
}
CluseterList clist = new CluseterList();
clist.addCluster(blist.getBisectList().get(lowest).getC1());
clist.addCluster(blist.getBisectList().get(lowest).getC2());
return clist.getClusters();
}
