BayesianWordAnalyzer.java

/*
 * Encog(tm) Java Examples v3.4
 * http://www.heatonresearch.com/encog/
 * https://github.com/encog/encog-java-examples
 *
 * Copyright 2008-2016 Heaton Research, Inc.
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 *   
 * For more information on Heaton Research copyrights, licenses 
 * and trademarks visit:
 * http://www.heatonresearch.com/copyright
 */
package org.encog.examples.ml.bayesian.words;

import java.util.ArrayList;
import java.util.List;

import org.encog.mathutil.probability.CalcProbability;
import org.encog.ml.bayesian.BayesianEvent;
import org.encog.ml.bayesian.BayesianNetwork;
import org.encog.ml.bayesian.EventType;
import org.encog.ml.bayesian.query.enumerate.EnumerationQuery;
import org.encog.util.Format;
import org.encog.util.text.BagOfWords;

public class BayesianWordAnalyzer {
	
	private int k;
	private BagOfWords classBag;
	private BagOfWords notClassBag;
	private BagOfWords totalBag;
	private final String className;
	private final String notClassName;
	private CalcProbability messageProbability;
	private String lastProblem;
	private final int classSampleCount;
	private final int notClassSampleCount;
	
	public BayesianWordAnalyzer(
			int theK, 
			String theClassName, 
			String[] classStrings, 
			String theNotClassName, 
			String[] notClassStrings)
	{
		this.k = theK;
		
		this.className = theClassName;
		this.notClassName = theNotClassName;
		
		this.classSampleCount = classStrings.length;
		this.notClassSampleCount = notClassStrings.length;
		
		this.classBag = new BagOfWords(this.k);
		this.notClassBag = new BagOfWords(this.k);
		this.totalBag = new BagOfWords(this.k);

		
		for(String line: classStrings) {
			this.classBag.process(line);
			totalBag.process(line);
		}
		
		for(String line: notClassStrings) {
			this.notClassBag.process(line);
			totalBag.process(line);
		}
		
		this.classBag.setLaplaceClasses(totalBag.getUniqueWords());
		this.notClassBag.setLaplaceClasses(totalBag.getUniqueWords());
		
		this.messageProbability = new CalcProbability(this.k);
		messageProbability.addClass(this.classSampleCount);
		messageProbability.addClass(this.notClassSampleCount);
	}

	
	public List<String> separateSpaces(String str) {
		List<String> result = new ArrayList<String>();
		StringBuilder word = new StringBuilder();

		for (int i = 0; i < str.length(); i++) {
			char ch = str.charAt(i);
			if (ch != '\'' && !Character.isLetterOrDigit(ch)) {
				if (word.length() > 0) {
					result.add(word.toString());
					word.setLength(0);
				}
			} else {
				word.append(ch);
			}
		}

		if (word.length() > 0) {
			result.add(word.toString());
		}
		
		return result;
	}
	
	public double probability(String m) {
		List<String> words = separateSpaces(m);
		
		BayesianNetwork network = new BayesianNetwork();
		BayesianEvent spamEvent = network.createEvent(this.className);
		
		int index = 0;
		for( String word: words) {
			BayesianEvent event = network.createEvent(word+index);
			network.createDependency(spamEvent, event);
			index++;
		}
		
		network.finalizeStructure();
		
		//SamplingQuery query = new SamplingQuery(network);
		EnumerationQuery query = new EnumerationQuery(network);
		
		double probSpam = messageProbability.calculate(0);

		spamEvent.getTable().addLine(probSpam, true);
		query.defineEventType(spamEvent, EventType.Outcome);
		query.setEventValue(spamEvent, true);
				
		index = 0;
		for( String word: words) {
			String word2 = word+index;
			BayesianEvent event = network.getEvent(word2);
			event.getTable().addLine(this.classBag.probability(word), true, true); // spam
			event.getTable().addLine(this.notClassBag.probability(word), true, false); // ham
			query.defineEventType(event, EventType.Evidence);
			query.setEventValue(event, true);
			index++;
		}

		//query.setSampleSize(100000000);
		query.execute();
		this.lastProblem = query.getProblem();
		//System.out.println(query.getProblem());
		return query.getProbability();		
	}


	/**
	 * @return the className
	 */
	public String getClassName() {
		return className;
	}


	/**
	 * @return the notClassName
	 */
	public String getNotClassName() {
		return notClassName;
	}
	
	public double getClassProbability() {
		return this.messageProbability.calculate(0);
	}
	
	public double getNotClassProbability() {
		return this.messageProbability.calculate(1);
	}
	
	public double probabilityWordClass(String word) {
		StringBuilder s = new StringBuilder();
		s.append("P(");
		s.append(word);
		s.append("|");
		s.append(this.className);
		s.append(")");
		this.lastProblem = s.toString();
		return this.classBag.probability(word);
	}
	
	public double probabilityWordNotClass(String word) {
		StringBuilder s = new StringBuilder();
		s.append("P(");
		s.append(word);
		s.append("|");
		s.append(this.notClassName);
		s.append(")");
		this.lastProblem = s.toString();
		return this.notClassBag.probability(word);
	}
	
	public String getLastProblem() {
		return this.lastProblem;
	}
	
}