两种实现模糊匹配的方法--python

以下举例以同一个excel中， sheet2的词语去匹配sheet1中词语找模糊匹配结果来举例
导入数据，读取excel中sheet1(被匹配的目标词库),sheet2(需要进行匹配的词)

import pandas as pd
import jieba
#需要进行匹配的词
attendee =  pd.read_excel('路径/testnn.xlsx',sheet_name='Sheet2')
#被匹配的目标词库
account =  pd.read_excel('路径/testnn.xlsx',sheet_name='Sheet1')
attendee = attendee.values
account = account.values
#print(attendee)
#print(account)

结果：

…………………………………………………………

一、分词匹配

把需要匹配的词语和目标词语做分词，对比分词匹配度判定关联关系

1、导入jieba分词包，对目标词和待匹配词进行分词，并将其导入至新字典中

#需要进行匹配的词的分词结果字典
Sheet2 = {}
for i in attendee:
    HCO=[]
    temp = jieba.cut(i[0], cut_all=False)
    for a in temp:
        HCO.append(a)
    Sheet2[i[0]] = HCO
#print(Sheet2)
#被匹配的目标词库的分词结果字典
Sheet1 = {}
for i in account:
     HCO = []
     temp = jieba.cut(i[0], cut_all=False)
     for a in temp:
         HCO.append(a)
     Sheet1[i[0]] = HCO
#print(Sheet1)

结果：

2、遍历分词后结果字典，对比相同的关键词并记录匹配情况

for i in Sheet1:
    a = i
    if i in Sheet2:
        #如果名称完全相同则返回名称
        resultstr = i
    #如果名称不完全相同，对比分词后的词语
    for j in Sheet2:
        b = j
        #需要进行匹配的词的分词数量
        origin_num = 0  
        #两分词结果中匹配成功的分词词语数量
        match_num = 0
        #存储需要进行匹配的词的分词结果
        origin_l=[]
        for k in Sheet1[i] :
            # xxxx代表分词结果中需要人工判定排除的异常词
            if k != 'xxxx':
                c = k
                origin_l.append(k)
                origin_num = origin_num +1
                target_l = []
                target_num = 0
                for h in Sheet2[j] :
                    # xxxx代表分词结果中需要人工判定排除的异常词
                    if h != 'xxxx':
                        d = h
                        target_num = target_num +1
                        target_l.append(h)
                        if c == d:
                            match_num = match_num + 1
        #选取符合条件的结果输出，每条词语对应一条结果
        if match_num > origin_num - match_num:
            data = {'origin_str': a, 'target_str': b, 'origin_l': origin_l, 'target_l': target_l,'origin_num': origin_num, 'target_num':target_num, 'match_num':match_num}
            print(data)

结果概览：

整体代码

import pandas as pd
import jieba
#需要进行匹配的词
attendee =  pd.read_excel('路径/testnn.xlsx',sheet_name='Sheet2')
#被匹配的目标词库
account =  pd.read_excel('路径/testnn.xlsx',sheet_name='Sheet1')
attendee = attendee.values
account = account.values
#print(attendee)
#print(account)

Sheet2 = {}
for i in attendee:
    HCO=[]
    temp = jieba.cut(i[0], cut_all=False)
    for a in temp:
        HCO.append(a)
    Sheet2[i[0]] = HCO
#print(Sheet2)
#被匹配的目标词库的分词结果字典
Sheet1 = {}
for i in account:
     HCO = []
     temp = jieba.cut(i[0], cut_all=False)
     for a in temp:
         HCO.append(a)
     Sheet1[i[0]] = HCO
#print(Sheet1)
for i in Sheet1:
    a = i
    if i in Sheet2:
        #如果名称完全相同则返回名称
        resultstr = i
    #如果名称不完全相同，对比分词后的词语
    for j in Sheet2:
        b = j
        #需要进行匹配的词的分词数量
        origin_num = 0
        #两分词结果中匹配成功的分词词语数量
        match_num = 0
        #存储需要进行匹配的词的分词结果
        origin_l=[]
        for k in Sheet1[i] :
            # xxxx代表分词结果中需要人工判定排除的异常词
            if k != 'xxxx':
                c = k
                origin_l.append(k)
                origin_num = origin_num +1
                target_l = []
                target_num = 0
                for h in Sheet2[j] :
                    # xxxx代表分词结果中需要人工判定排除的异常词
                    if h != 'xxxx':
                        d = h
                        target_num = target_num +1
                        target_l.append(h)
                        if c == d:
                            match_num = match_num + 1
        #选取符合条件的结果输出
        if match_num > origin_num - match_num:
            data = {'origin_str': a, 'target_str': b, 'origin_l': origin_l, 'target_l': target_l,'origin_num': origin_num, 'target_num':target_num, 'match_num':match_num}
            print(data)

二、距离匹配

调用fuzzywuzzy包中直接进行判断，采用距离匹配方式
两个字符串之间，由一个转成另一个所需的最少编辑操作次数。
编辑操作包括：将一个字符替换成另一个字符，插入字符，删除字符。
一般来说，编辑距离越小，两个串的相似度越大
整体代码

import pandas as pd
import jieba
from fuzzywuzzy import fuzz
from fuzzywuzzy import process

attendee =  pd.read_excel('路径/testnn.xlsx',sheet_name='Sheet2')
account =  pd.read_excel('路径/testnn.xlsx',sheet_name='Sheet1')
attendee = attendee.values
account = account.values


Sheet2 = {}
for i in attendee:
    HCO=[]
    temp = jieba.cut(i[0], cut_all=False)
    for a in temp:
        HCO.append(a)
    Sheet2[i[0]] = HCO
print(Sheet2)
Sheet1 = {}
for i in account:
     HCO = []
     temp = jieba.cut(i[0], cut_all=False)
     for a in temp:
         HCO.append(a)
     Sheet1[i[0]] = HCO
print(Sheet1)
target_l = []
data = []
n = 0
for j in Sheet2:
    target_l.append(j)
for i in Sheet1:
    n = n+1
    target= {'搜索公司':i,'目标公司': process.extractOne( i, target_l )[0],'目标权重': process.extractOne( i, target_l )[1]}
    data.append(target)
    print (data)
df1 = pd.DataFrame(data)
print(df1)

writer = pd.ExcelWriter('路径/testmm.xlsx')
df1.to_excel(writer, 'Final')
writer.save()
writer.close()

结果概览：