字符串相似度

由于工作需要,公司数据要与客户的数据进度匹配,在网上找了一些方法感觉还不错,我采用的是余弦算法。这是我的参考地址http://www.maoguangpu.com/post/117.html。 里面只做对中文的处理,我做了些改进,中,英,符号都可进行。上代码

import java.io.UnsupportedEncodingException;
import java.util.HashMap;
import java.util.Iterator;
import java.util.Map;

public class StringSimilar {
    
    
    public static void main (String [] args){
        System.out.println(StringSimilar.getSimilarity("a中国", "国a"));
    }
    public static double getSimilarity(String doc1, String doc2) {
        doc1=doc1.trim();
        doc2=doc2.trim();
        if (doc1 != null && doc1.length() > 0 && doc2 != null
                && doc2.length() > 0) {
            
            Map<Integer, int[]> AlgorithmMap = new HashMap<Integer, int[]>();
            
            //将两个字符串中的中文字符以及出现的总数封装到,AlgorithmMap中
            for (int i = 0; i < doc1.length(); i++) {
                char d1 = doc1.charAt(i);
                if(isHanZi(d1)){
                    int charIndex = getGB2312Id(d1);
                    if(charIndex != -1){
                        int[] fq = AlgorithmMap.get(charIndex);
                        if(fq != null && fq.length == 2){
                            fq[0]++;
                        }else {
                            fq = new int[2];
                            fq[0] = 1;
                            fq[1] = 0;
                            AlgorithmMap.put(charIndex, fq);
                        }
                    }
                }else{
                    int by=(int)-String.valueOf(d1).getBytes()[0];
                    int[] fq = AlgorithmMap.get(by);
                    if(fq != null && fq.length == 2){
                        fq[0]++;
                    }else {
                        fq = new int[2];
                        fq[0] = 1;
                        fq[1] = 0;
                        AlgorithmMap.put(by, fq);
                    }
                }
            }

            for (int i = 0; i < doc2.length(); i++) {
                char d2 = doc2.charAt(i);
                if(isHanZi(d2)){
                    int charIndex = getGB2312Id(d2);
                    if(charIndex != -1){
                        int[] fq = AlgorithmMap.get(charIndex);
                        if(fq != null && fq.length == 2){
                            fq[1]++;
                        }else {
                            fq = new int[2];
                            fq[0] = 0;
                            fq[1] = 1;
                            AlgorithmMap.put(charIndex, fq);
                        }
                    }
                }else{
                    int by=(int)-String.valueOf(d2).getBytes()[0];
                    int[] fq = AlgorithmMap.get(by);
                    if(fq != null && fq.length == 2){
                        fq[1]++;
                    }else {
                        fq = new int[2];
                        fq[0] = 0;
                        fq[1] = 1;
                        AlgorithmMap.put(by, fq);
                    }
                }
            }
            
            Iterator<Integer> iterator = AlgorithmMap.keySet().iterator();
            double sqdoc1 = 0;
            double sqdoc2 = 0;
            double denominator = 0; 
            while(iterator.hasNext()){
                int[] c = AlgorithmMap.get(iterator.next());
                denominator += c[0]*c[1];
                sqdoc1 += c[0]*c[0];
                sqdoc2 += c[1]*c[1];
            }
            
            return denominator / Math.sqrt(sqdoc1*sqdoc2);
        } else {
            throw new NullPointerException(
                    " the Document is null or have not cahrs!!");
        }
    }

    public static boolean isHanZi(char ch) {
        // 判断是否汉字
        return (ch >= 0x4E00 && ch <= 0x9FA5);

    }

    /**
     * 根据输入的Unicode字符,获取它的GB2312编码或者ascii编码,
     * 
     * @param ch
     *            输入的GB2312中文字符或者ASCII字符(128个)
     * @return ch在GB2312中的位置,-1表示该字符不认识
     */
    public static short getGB2312Id(char ch) {
        try {
            byte[] buffer = Character.toString(ch).getBytes("GB2312");
            if (buffer.length != 2) {
                // 正常情况下buffer应该是两个字节,否则说明ch不属于GB2312编码,故返回'?',此时说明不认识该字符
                
                return -1;
            }
            int b0 = (int) (buffer[0] & 0x0FF) - 161; // 编码从A1开始,因此减去0xA1=161
            int b1 = (int) (buffer[1] & 0x0FF) - 161; // 第一个字符和最后一个字符没有汉字,因此每个区只收16*6-2=94个汉字
            return (short) (b0 * 94 + b1);
        } catch (UnsupportedEncodingException e) {
            e.printStackTrace();
        }
        return -1;
    }
}

点赞