interfieldpipe.java

常用机器学习算法,java编写源代码,内含常用分类算法,包括说明文档

package edu.umass.cs.mallet.projects.seg_plus_coref.coreference;

import com.wcohen.secondstring.*;
import edu.umass.cs.mallet.base.types.*;
import edu.umass.cs.mallet.base.classify.*;
import edu.umass.cs.mallet.base.pipe.*;
import edu.umass.cs.mallet.base.pipe.iterator.*;
import edu.umass.cs.mallet.base.util.*;
import java.util.*;
import java.lang.*;
import java.io.*;

public class InterFieldPipe extends Pipe
{

	public InterFieldPipe () {}

	public Instance pipe (Instance carrier) {
		NodePair pair = (NodePair)carrier.getData();

		Citation s1 = (Citation)pair.getObject1();
		Citation s2 = (Citation)pair.getObject2();

		double authorVal = 0.0;
		double titleVal = 0.0;
		double paperVal = 0.0;
		double dateVal = 0.0;
		double venueVal = 0.0;
		authorVal = pair.getFeatureValue("trigramTFIDF_author");
		titleVal = pair.getFeatureValue("trigramTFIDF_title");
		paperVal = pair.getFeatureValue("SamePaperType");
		dateVal = pair.getFeatureValue("Same_year");
		venueVal = pair.getFeatureValue("trigramTFIDF_venue");

		//System.out.println("venue value in InterField: " + venueVal);

		if (authorVal > 0.8 && dateVal > 0.8 && titleVal > 0.9 && paperVal > 0.9 && venueVal > 0.9)
			pair.setFeatureValue("GeneralAgreement1", 1.0);

		/*
		if (((!s1.getField(Citation.tech).equals("") && s2.getField(Citation.tech).equals("")) ||
				 (!s2.getField(Citation.tech).equals("") && s1.getField(Citation.tech).equals(""))) &&
				(pair.getFeatureValue("TitleSimilarity") > 0.9) )
		{
		
			System.out.println("TitleSimilarOnlyOneHasTech --> " +
			s1.getBaseString() + " :: " + s2.getBaseString());
			pair.setFeatureValue("TitleSimilarOnlyOneHasTech", 1.0);
		}

		if (((!s1.getField(Citation.booktitle).equals("") && s2.getField(Citation.booktitle).equals("")) ||
				(!s2.getField(Citation.booktitle).equals("") && s1.getField(Citation.booktitle).equals(""))) &&
				(pair.getFeatureValue("TitleSimilarity") > 0.9) )
		{
			pair.setFeatureValue("TitleSimilarOnlyOneHasBookTitle", 1.0);
		}
		
		if (((!s1.getField(Citation.journal).equals("") && s2.getField(Citation.journal).equals("")) ||
				 (!s2.getField(Citation.journal).equals("") && s1.getField(Citation.journal).equals(""))) &&
				(pair.getFeatureValue("TitleSimilarity") > 0.9) )
		{
			pair.setFeatureValue("TitleSimilarOnlyOneHasJournal", 1.0);
		}
		double v1 = ((double)pair.getFeatureValue("TitleSimilarity") *
								 (double)pair.getFeatureValue("AuthorOverlap"));
								 
		if (v1 > 0)
			pair.setFeatureValue("AuthorANDTitle", v1);

	*/
		return carrier;
	}
}