Python基于textdistance实现计算文本相似度

pip install textdistance

# 使用扩展库，提高性能
pip install "textdistance[extras]"

import textdistance

# 计算编辑距离
distance = textdistance.levenshtein.distance("如何更换花呗绑定银行卡", "花呗更改绑定银行卡")
print("编辑距离:", distance)

# 计算余弦相似度
similarity = textdistance.cosine.similarity("如何更换花呗绑定银行卡", "花呗更改绑定银行卡")
print("余弦相似度:", similarity)

# 计算 Jaccard 系数
coefficient = textdistance.jaccard("如何更换花呗绑定银行卡", "花呗更改绑定银行卡")
print("Jaccard 系数:", coefficient)

# 计算 Hamming 距离
distance = textdistance.hamming.distance("如何更换花呗绑定银行卡", "花呗更改绑定银行卡")
print("Hamming 距离:", distance)

import textdistance

# 拼写检查
word = "发愤图强"
possible_spellings = ["发奋图强", "发奋图", "发愤图"]

for spelling in possible_spellings:
    distance = textdistance.levenshtein.distance(word, spelling)
    if distance <= 1:
        print("可能的正确拼写:", spelling)

import textdistance

# 文档相似度计算
doc1 = "Python is a programming language"
doc2 = "Python is used for web development"
doc3 = "Java is a programming language"

similarity1 = textdistance.cosine.similarity(doc1, doc2)
similarity2 = textdistance.cosine.similarity(doc1, doc3)

print("文档1和文档2的余弦相似度:", similarity1)
print("文档1和文档3的余弦相似度:", similarity2)

import textdistance

# 数据清洗
data = ["配偶", "原配", "元配", "老婆", "夫人", "爱人"]

for i in range(len(data)):
    for j in range(i+1, len(data)):
        distance = textdistance.hamming.distance(data[i], data[j])
        if distance <= 1:
            print("相似但不完全相同的数据:", data[i], data[j])

import textdistance

# 姓名匹配
name1 = "李建国"
name2 = "张建国"
name3 = "王爱国"

coefficient1 = textdistance.jaccard.normalized_similarity(name1, name2)
coefficient2 = textdistance.jaccard.normalized_similarity(name1, name3)

print("姓名1和姓名2的Jaccard系数:", coefficient1)
print("姓名1和姓名3的Jaccard系数:", coefficient2)