计算tfidf，关键词抽取---python

1、读入文本内容

corpos = pandas.DataFrame(columns=[‘filePath‘,‘content‘])
for root ,dirs,files in os.walk(r‘H:\19113117 - 副本‘):
    for name in files:
        filePath=root+‘\\‘+name
        f = codecs.open(filePath,‘r‘,‘utf-8‘)
        content=f.read()
        f.close()
        corpos.loc[len(corpos)+1]=[filePath,content.strip()]

2、将手动分完词的文本进行词频统计

filePaths=[]
segments=[]
for filePath,content in corpos.itertuples(index=False):
    for item in content.split(‘/‘):
        segments.append(item)
        filePaths.append(filePath)
segmentDF=pandas.DataFrame({‘filePath‘:filePaths,‘segments‘:segments})

segStat = segmentDF.groupby(
            by=["filePath","segments"]
        )["segments"].agg({
            "计数":numpy.size
        }).reset_index();

3、计算tf值

textVector=segStat.pivot_table(
           index=‘segments‘,
           values=‘计数‘,
           columns=‘filePath‘,
           fill_value=0)
tF=(1+numpy.log(textVector)).as_matrix()

4、计算IDF

def handle(x):
    idf=1+numpy.log(len(corpos)/(numpy.sum(x>0)+1))
    return idf
zhuan=textVector.T
iDF=zhuan.apply(handle).as_matrix()
iDF=iDF.reshape(8889,1)

5、计算tfidf

TFIDF=tF*iDF
tFIDF_DF=pandas.DataFrame(TFIDF)

6、将每个文本中tfidf值排名前100的词和相应的tfidf值输出

file=[]
for root ,dirs,files in os.walk(r‘H:\19113117 - 副本‘):
    for name in files:
        name=name[0:-4]
        file.append(name)
for i in range(len(corpos)):
    sort=pandas.DataFrame(tFIDF_DF.loc[:,i].order(ascending=False)[:100]).reset_index()
    names = sort.columns.tolist()
    names[names.index(i)] = ‘value‘
    sort.columns = names
    tagis = textVector.index[sort.index]
    print(file[i])
    for t in range(len(tagis)):
        print(tagis[t],sort.loc[t].value)