全文检索几种词向量模型

蔚落 2024-04-19 13:07 4阅读 0赞

1、倒排索引模型  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70]  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 1]![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 2]![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 3]  
2、布尔检索类型  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 4]  
3、TF-IDF权重计算  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 5]![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 6]  
下面是TF-IDF的JAVA代码实现。

public class TFIDF {
        public double tf(List<String> doc,String term){
            double termFrequency = 0;
            for (String str : doc) {
                if(str.equalsIgnoreCase(term)){
                    termFrequency ++ ;
                }
            }
            return termFrequency / doc.size();
        }
        public double df(List<List<String>> docs,String term){
            int n = 0;
            if(term != null && term != ""){
                for (List<String> doc : docs) {
                    for (String word : doc) {
                        if (term.equalsIgnoreCase(word)){
                            n ++;
                            break;
                        }
                    }
                }
            }else {
                System.out.println("term不能为null或空串");
            }
            return n;
        }
        public double idf(List<List<String>> docs,String term){
            return Math.log(docs.size()/(double)df(docs,term)+1);
        }
        public double tfidf(List<String> doc,List<List<String>> docs,String term){
            return tf(doc,term) * idf(docs,term);
        }
    
        public static void main(String args[]){
            List<String> doc1 = Arrays.asList("人工","智能","成为","互联网","大会","焦点");
            List<String> doc2 = Arrays.asList("谷歌","推出","开源","人工","智能","系统","工具");
            List<String> doc3 = Arrays.asList("互联网","的","未来","在","人工","智能");
            List<String> doc4 = Arrays.asList("谷歌","开源","机器","学习","工具");
    
            List<List<String>> documents = Arrays.asList(doc1,doc2,doc3,doc4);
            TFIDF calculator = new TFIDF();
            System.out.println(calculator.tf(doc2,"谷歌"));
            System.out.println(calculator.df(documents,"谷歌"));
            double tfidf = calculator.tfidf(doc2, documents, "谷歌");
            System.out.println("TF-IDF(谷歌) = " + tfidf);
        }
    }
    
    结果：
    0.14285714285714285
    2.0
    TF-IDF(谷歌) = 0.15694461266687282

4、向量空间模型  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 7]![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 8]  
下面是相似度的JAVA代码实现

public class Vsm {
        public static double calCosSim(Map<String,Double> v1,Map<String,Double> v2){
            double sclar = 0.0,norm1 = 0.0,norm2 = 0.0,similarity = 0.0;
            Set<String> v1Keys = v1.keySet();
            Set<String> v2Keys = v2.keySet();
            Set<String> both = new HashSet<>();
            both.addAll(v1Keys);
            both.retainAll(v2Keys);
            System.out.println(both);
            for (String str1 : both) {
                sclar += v1.get(str1) * v2.get(str1);
            }
            for (String str1 : v1.keySet()) {
                norm1+= Math.pow(v1.get(str1),2);
            }
            for (String str2 : v2.keySet()) {
                norm2+= Math.pow(v2.get(str2),2);
            }
            similarity = sclar / Math.sqrt(norm1*norm2);
            System.out.println("sclar: "+sclar);
            System.out.println("norm1: "+norm1);
            System.out.println("norm2: "+norm2);
            System.out.println("similarity : "+similarity);
            return similarity;
        }
    
        public static void main(String[] args) {
            Map<String,Double> m1 = new HashMap<>();
            m1.put("Hello",1.0);
            m1.put("css",2.0);
            m1.put("Lucene",3.0);
    
            Map<String,Double> m2 = new HashMap<>();
            m2.put("Hello",1.0);
            m2.put("Word",2.0);
            m2.put("Hadoop",3.0);
            m2.put("java",4.0);
            m2.put("html",1.0);
            m2.put("css",2.0);
            calCosSim(m1,m2);
        }
    }
    
    
    结果：
    
    [css, Hello]
    sclar: 5.0
    norm1: 14.0
    norm2: 35.0
    similarity : 0.22587697572631282

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/18/b4684efae13545beb62a998205df5301.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 1]: https://img-blog.csdnimg.cn/2019091816530283.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 2]: https://img-blog.csdnimg.cn/20190918165329104.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 3]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/18/c4a8d28e50de438e94d5fd7129b1e96e.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 4]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/18/a997e7f9eb6f41a898508e59f7a26bbe.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 5]: https://img-blog.csdnimg.cn/20190918165547787.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 6]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/18/93c3e2b2f3434334b773cba2d6549957.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 7]: https://img-blog.csdnimg.cn/20190918172057798.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3lhbmdhbmcxMjIz_size_16_color_FFFFFF_t_70 8]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/18/bfa12f60b8824623adc20395f94f091b.png