/**
* Returns the list of only points excluding the class label and point ID
*
* @param data
* @return
*/
public List<String[]> findPoints(DataCluster data) {
List<String[]> list = new ArrayList<String[]>();
for (int i = 0; i < data.sizeOfRecords(); i++) {
list.add(getPoints(data.getAttributes(), data.getRecordAt(i)));
}
return list;
}
/**
* Adds cluster points from data table
*
* @param list of clusters
* @param data
*/
public void addClusterPoints(CluseterList list, DataCluster data) {
list.clearsClusterPoints();
for (int i = 0; i < data.sizeOfRecords(); i++) {
int temp =
findClosestClusterAndAddPoint(list, getPoints(data.getAttributes(), data.getRecordAt(i)));
list.getClusterAt(temp).addPoints(data.getRecordAt(i));
}
}
/**
* Normalizes data in a cluster
*
* @param data
* @return
*/
public DataCluster normalizeData(DataCluster data) {
DataCluster temp = new DataCluster();
temp.setAttributes(data.getAttributes());
List<Double> max = new ArrayList<Double>();
List<Double> min = new ArrayList<Double>();
for (int i = 0; i < data.getAttributes().size(); i++) {
if (data.getAttributesat(i).getType() == TypeAttribute.continuous) {
List<Double> listDouble = data.getColumnAt(i);
max.add(Collections.max(listDouble));
min.add(Collections.min(listDouble));
}
}
for (int k = 0; k < data.sizeOfRecords(); k++) {
RecordCluster rec = new RecordCluster();
// List<String> str = new ArrayList<>();
int mark = 0;
for (int i = 0; i < data.numberOfAttributes(); i++) {
if (data.getAttributes().get(i).getType() == TypeAttribute.continuous) {
rec.addValue(
new ErrorsAndGain()
.roundOff(
((Double.parseDouble(data.getRecordAt(k).getValueat(i)) - min.get(mark))
/ (max.get(mark) - min.get(mark))),
4));
mark++;
} else {
rec.addValue(data.getRecordAt(k).getValueat(i));
}
}
temp.addRecord(rec);
}
return temp;
}