相似度算法：余弦定理

最新推荐文章于 2025-11-10 02:03:11 发布

原创最新推荐文章于 2025-11-10 02:03:11 发布 · 923 阅读

1 ·

CC 4.0 BY-SA版权

文章标签：

#算法 #java #string

数据结构和算法专栏收录该内容

14 篇文章

订阅专栏

本文介绍了一种使用Java实现的字符串相似度计算方法，通过构建字符向量映射并利用余弦相似度公式来衡量两个字符串之间的相似程度。该方法首先统计每个字符在两个字符串中出现的频次，接着计算这些频次向量之间的余弦相似度。

求两个字符串的相似度，直接贴代码：

public class Similarity {
        LinkedHashMap<Character, int[]> vectorMap = new LinkedHashMap<Character, int[]>();  

        int[] tempArray = null;  

        public Similarity(String string1, String string2) {  
            for (Character character1 : string1.toCharArray()) {  
                if (vectorMap.containsKey(character1)) {  
                    vectorMap.get(character1)[0]++;  
                } else {  
                    tempArray = new int[2];  
                    tempArray[0] = 1;  
                    tempArray[1] = 0;  
                    vectorMap.put(character1, tempArray);  
                }  
            }  
            for (Character character2 : string2.toCharArray()) {  
                if (vectorMap.containsKey(character2)) {  
                    vectorMap.get(character2)[1]++;  
                } else {  
                    tempArray = new int[2];  
                    tempArray[0] = 0;  
                    tempArray[1] = 1;  
                    vectorMap.put(character2, tempArray);  
                }  
            } 
            for (Map.Entry<Character, int[]> entry : vectorMap.entrySet()) {  
                System.out.println("Key = " + entry.getKey() + ", Value = " + entry.getValue()[0] +","+entry.getValue()[1]); 
            } 
        }  

        // 求余弦相似度  
        public double sim() {  
            double result = 0;  
            //cos c =  a*b/(|a|*|b|)
            result = pointMulti(vectorMap) / sqrtMulti(vectorMap);  
            return result;  
        }  

        private double sqrtMulti(Map<Character, int[]> paramMap) {  
            double result = 0;  
            result = squares(paramMap); 
            result = Math.sqrt(result);
            System.out.println("sqrtMulti result:"+result);
            return result;  
        }  

        // 点乘法  
        private double pointMulti(Map<Character, int[]> paramMap) {  
            double result = 0;  
            Set<Character> keySet = paramMap.keySet();  
            for (Character character : keySet) {  
                int temp[] = paramMap.get(character);  
                result += (temp[0] * temp[1]);  
            }  
            System.out.println("pointMulti result:"+result);
            return result;  
        }  

        // 求平方和  
        private double squares(Map<Character, int[]> paramMap) {  
            double result1 = 0.00;  //向量1的模
            double result2 = 0.00;  //向量2的模
            double resultproduct = 0.00;//向量积
            Set<Character> keySet = paramMap.keySet();  
            for (Character character : keySet) {  
                int temp[] = paramMap.get(character);  
                result1 += (temp[0] * temp[0]);  
                result2 += (temp[1] * temp[1]);  
            } 
            resultproduct = result1 * result2;
            return resultproduct;  
        }  
        public static void main(String[] args) {  
            String s1 = "我不是一个帅哥";  
            String s2 = "一个帅哥是我";  
            Similarity similarity1 = new Similarity(s1, s2);    
            System.out.println(similarity1.sim());  
        }  
}