TF-IDF 词袋

最新推荐文章于 2024-07-18 18:59:32 发布

张一爻

最新推荐文章于 2024-07-18 18:59:32 发布

阅读量156

点赞数

本文链接：https://blog.csdn.net/weixin_43069769/article/details/106772350

版权

def TF_IDF(docList):
    result = dict()
    m = len(docList)
    WFD1 = 0
    distion_word = set(drop_2_to_1_dim(docList))
    leve1,leve2 = [],[]
    for doc in docList:
        n = len(doc)
        tmpTF = np.array([(doc==word).sum()/n for word in distion_word])
        leve1.append(tmpTF)
        tmpIDF = np.array([word in doc for word in distion_word]).astype(int)
        leve2.append(tmpIDF)
    TF = np.array(leve1)
    count_word = np.array(leve2).T.dot(np.ones(m))
    IDF = np.log(m/(count_word+1))
    return TF*IDF