TF-IDF算法——原理及实现

package com.jsptpd.wordpart;

import java.util.Arrays;
import java.util.List;

/**
 * //TF-IDF算法——原理及实现
 *
 */


public class App 
{
	/**
	 * 词频统计
	 */
	public double  tf(Listdoc,String item) {
		double termFrequency = 0;
		for(String str:doc) {
			if(str.equalsIgnoreCase(item)) {
			termFrequency++;
			}
		}
		
		return termFrequency;
		
	}
	
	/***
	 * 文档频率统计
	 */
	public int df(List<List> docs,String item) {
		int n =0;
		if(item != null  && item != "") {
			for(Listdoc:docs) {
				for(String word:doc) {
					if(word.equalsIgnoreCase(item)) {
						n++;
						break;
					}
				}
			}
			
		}else {
			System.out.println("item 不能为null或者空串");
		}
		
		return n;
	}
	
	/**
	 * 逆文档频率
	 */
	public double idf(List<List> docs,String item) {
		return Math.log(docs.size()/(double) df(docs,item)+1);
	}
	
	/*
	 * 词频
	 */
	public double tfIdf(Listdoc,List<List> docs,String item) {
		return tf(doc,item)*idf(docs,item);
		
	}
	
	
	
	
    public static void main( String[] args )
    {
    	
        
    	Listdoc1 = Arrays.asList("人工","智能","成为","互联网","大会","焦点");
    	Listdoc2 = Arrays.asList("谷歌","推出","开源","人工","智能","系统","工具");
    	Listdoc3 = Arrays.asList("互联网","的","未来","在","人工","智能");
    	Listdoc4 = Arrays.asList("谷歌","开源","机器","学习","工具");
    	List<List> documents = Arrays.asList(doc1,doc2,doc3,doc4);
    	App app1 = new App();
    	;
    	System.out.println(app1.tf(doc2, "谷歌"));
    	System.out.println(app1.df(documents, "谷歌"));
    	System.out.println(app1.tfIdf(doc4,documents, "学习"));
    }
}