python 相似度算法 距离算法

更多算法可查看，包含很多距离算法讲解：机器学习中的数学——距离定义（五）：标准化的欧几里得距离（Standardized Euclidean Distance）_von Neumann的博客-CSDN博客_标准化欧几里得距离
#相似度算法
class Similarity():
    def __init__(self):
        pass
    # 欧式距离， (x-y)平方累加在开放 参考：https://blog.csdn.net/hy592070616/article/details/121461909?spm=1001.2014.3001.5501
    def Euclidean_distance(slef,vector1,vector2):
        count = 0
        for i, t in enumerate(vector1):
            count += (vector1[i] - vector2[i]) ** 2
        print('Euclidean_distance', 1 / ((count ** 0.5) + 1))
        return 1 / ((count ** 0.5) + 1)
    # 余弦距离，参考：https://blog.csdn.net/hy592070616/article/details/122271927?spm=1001.2014.3001.5501
    def Cosine_distance(self,vector1,vector2):
        count, d = 0, 0
        for i, t in enumerate(vector1):
            d += (vector1[i] * vector2[i])
            count += (vector1[i] ** 2) * (vector2[i] ** 2)
        print('Cosine_distance', d / (count+1)) #余弦取值范围为[-1,1]  0以上正相关，0一下负相关，0时两向量垂直
        return d / (count+1)
    # 曼哈顿距离 参考：https://blog.csdn.net/hy592070616/article/details/121569933?spm=1001.2014.3001.5501
    def Manhattan_distance(self,vector1,vector2):
        count = 0
        for i, t in enumerate(vector1):
            count += abs(vector1[i] - vector2[i]) #abs() 绝对值函数
        print('Manhattan_distance', 1 / (count + 1))
        return 1 / (count + 1)
    # 杰卡德距离 参考：https://blog.csdn.net/qq_21484461/article/details/125570951
    def Jaccard_distance(slef, vector1,vector2):
        word_bag = []
        # 交集
        for i in vector1:
            if i in vector2:
                word_bag.append(i)
        print(set(word_bag))
        # 并集
        ret = set(vector1) | set(vector2) #并不会有重复值出现
        print(ret)
        print('Jaccard_distance', ((len(ret) - len(word_bag)) / len(ret)))
        return (len(ret) - len(word_bag)) / len(ret)
    # 汉明距离
    def Hamming_distance(self,vector1,vector2):
        count = 0
        for x,y in zip(vector1,vector2):
            if x != y:
                count += 1
        print('Hamming_distance',count/len(vector1))
        return count/len(vector1)
    # 编辑距离
    def EditDistance(self,x, y):
        import numpy as np
        dp = np.zeros((len(x) + 1, len(y) + 1))
        for i in range(len(x) + 1):
            dp[i][0] = i
        for j in range(len(y) + 1):
            dp[0][j] = j

        for i in range(1, len(x) + 1):
            for j in range(1, len(y) + 1):
                delta = 0 if x[i - 1] == y[j - 1] else 1
                dp[i][j] = min(dp[i - 1][j - 1] + delta, min(dp[i - 1][j] + 1, dp[i][j - 1] + 1))
        print(int(dp[len(x)][len(y)]))
        return int(dp[len(x)][len(y)])

#
# s = Similarity()
# a = ['a','c','s','a']
# b = ['a','d','b','d']
# s.Jaccard_distance(a,b)
# s.Hamming_distance(a,b)
# s.EditDistance(a,b)