prerna.algorithm.learning.util.NumericalCluster Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of semoss Show documentation
SEMOSS
The newest version!
package prerna.algorithm.learning.util;

import java.util.HashMap;
import java.util.HashSet;
import java.util.Hashtable;
import java.util.List;
import java.util.Map;
import java.util.Set;

public class NumericalCluster {

	private Set attributes = new HashSet();
	
	private Map distanceMeasureForAttribute;
	private Map weights = new HashMap();

	/**
	 * Default constructor
	 */
	public NumericalCluster(Map w) {
		this.distanceMeasureForAttribute = new Hashtable();
		this.weights = w;
		
		this.setAttributes(distanceMeasureForAttribute.keySet());
	}
	
	public Double getSimilarity(List attributeName, List value, int indexToSkip) {
		double similarity = 0.0;
		int numAttr = attributeName.size();
		for(int i = 0; i < numAttr; i++) {
			if(i==indexToSkip) {
				continue;
			}
			String attribute = attributeName.get(i);
			Double v = value.get(i);
			Double weight = weights.get(attribute);
			if(v==null || v.isNaN()) {
				v = distanceMeasureForAttribute.get(attribute).getNullRatio();
				similarity = similarity + v*weight;
				continue;
			}
			Double center = distanceMeasureForAttribute.get(attribute).getCentroidValue();
			//using euclidean distance
			similarity = similarity + (Math.pow(v-center, 2))*weight;
		}
		
		similarity = 1 - Math.sqrt(similarity);
		
		return similarity;
	}
	
	public Double getSimilarity(String attribute, Double value) {
		Double weight = weights.get(attribute);
		if(value==null) {
			value = distanceMeasureForAttribute.get(attribute).getNullRatio();
			return value*weight;
		}
		Double center = distanceMeasureForAttribute.get(attribute).getCentroidValue();
		//using euclidean distance
		return 1 - Math.sqrt((Math.pow(value-center, 2))*weight);
	}
	
	public void partialAddToCluster(List attributeName, List value, double factor) {
		for(int i = 0; i < attributeName.size(); i++) {
			this.addToCluster(attributeName.get(i), value.get(i), factor);
		}
	}
	
	public void addToCluster(String attributeName, Double value, double factor) {
		IClusterDistanceMode distanceMeasure = distanceMeasureForAttribute.get(attributeName);
		distanceMeasure.addPartialToCentroidValue(value, factor);
	}

	public void partialRemoveFromCluster(List attributeName, List value, double factor) {
		for(int i = 0; i < attributeName.size(); i++) {
			this.removeFromCluster(attributeName.get(i), value.get(i), factor);
		}	
	}
	
	public void removeFromCluster(String attributeName, Double value, double factor) {
		IClusterDistanceMode distanceMeasure = distanceMeasureForAttribute.get(attributeName);
		distanceMeasure.removePartialFromCentroidValue(value, factor);
	}

	
	public void addToCluster(List attributeName, List value) {
		for(int i = 0; i < attributeName.size(); i++) {
			this.addToCluster(attributeName.get(i), value.get(i));
		}
	}
	
	public void addToCluster(String attributeName, Double value) {
		IClusterDistanceMode distanceMeasure = distanceMeasureForAttribute.get(attributeName);
		distanceMeasure.addToCentroidValue(value);
	}

	public void removeFromCluster(List attributeName, List value) {
		for(int i = 0; i < attributeName.size(); i++) {
			this.removeFromCluster(attributeName.get(i), value.get(i));
		}	
	}
	
	public void removeFromCluster(String attributeName, Double value) {
		IClusterDistanceMode distanceMeasure = distanceMeasureForAttribute.get(attributeName);
		distanceMeasure.removeFromCentroidValue(value);
	}

	public double getClusterSimilarity(NumericalCluster c2, String instanceType) {
		double similarity = 0;
		if(c2.isEmpty() || this.isEmpty()) {
			return similarity;
		}
		
		for(String attributeName : this.distanceMeasureForAttribute.keySet()) {
			if(attributeName.equals(instanceType)) {
				continue;
			}
			Double weight = weights.get(attributeName);

			Double thisCenterValue = this.getCenterValueForAttribute(attributeName);
			Double centerValue = c2.getCenterValueForAttribute(attributeName);
			if(thisCenterValue == null || thisCenterValue.isNaN() || centerValue == null || centerValue.isNaN()) {
				double thisNumNull = this.getCenterNumNullsForAttribute(attributeName);
				double thisNumVals = this.getCenterNumInstancesForAttribute(attributeName);
				double numNull = c2.getCenterNumNullsForAttribute(attributeName);
				double numVals = c2.getCenterNumInstancesForAttribute(attributeName);
				similarity = similarity + (thisNumNull + numNull) / (thisNumVals + numVals + thisNumNull + numNull) * weight;
			} else {
				//using euclidean distance
				similarity = similarity + (Math.pow(thisCenterValue - centerValue, 2))*weight;
			}
		}
		
		similarity = (1 - Math.sqrt(similarity));
		return similarity;
	}

	public Map getWeights() {
		return this.weights;
	}
	
	public void setDistanceMode(String attributeName, IClusterDistanceMode distanceMeasure) {
		distanceMeasureForAttribute.put(attributeName, distanceMeasure);
	}
	
	public void reset() {
		for(String key: distanceMeasureForAttribute.keySet()) {
			distanceMeasureForAttribute.get(key).reset();
		}
	}

	public boolean isEmpty() {
		return distanceMeasureForAttribute.isEmpty();
	}
	
	//////////////START METHODS TO EXTRACT INFORMATION FROM DISTANCE MEASURE//////////////
	public Double getCenterValueForAttribute(String attributeName) {
		return this.distanceMeasureForAttribute.get(attributeName).getCentroidValue();
	}
	
	public double getCenterNumInstancesForAttribute(String attributeName) {
		return this.distanceMeasureForAttribute.get(attributeName).getNumInstances();
	}
	
	public double getCenterNumNullsForAttribute(String attributeName) {
		return this.distanceMeasureForAttribute.get(attributeName).getNumNull();
	}
	
	public double getPreviousCenterValueForAttribute(String attributeName) {
		return this.distanceMeasureForAttribute.get(attributeName).getPreviousCentroidValue();
	}
	
	public double getChangeToCenterForAttribute(String attributeName) {
		return this.distanceMeasureForAttribute.get(attributeName).getChangeToCentroidValue();
	}
	
	public boolean getIsPreviousNull(String attributeName) {
		return this.distanceMeasureForAttribute.get(attributeName).isPreviousNull();
	}
	//////////////END METHODS TO EXTRACT INFORMATION FROM DISTANCE MEASURE//////////////

	/**
	 * @return the attributes
	 */
	public Set getAttributes() {
		return attributes;
	}

	/**
	 * @param attributes the attributes to set
	 */
	public void setAttributes(Set attributes) {
		this.attributes = attributes;
	}
	
}