基于加权词向量和卷积神经网络的新闻文本分类  

News Text Classification Based on Weighted Word Vector and CNN

作　　者：胡万亭 贾真[2] HU Wan-Ting;JIA Zhen(Puyang Institute of Technology,Henan University,Puyang 457000,China;School of Information Science and Technology,Southwest Jiaotong University,Chengdu 611756,China)

机构地区：[1]河南大学濮阳工学院,濮阳457000 [2]西南交通大学信息科学与技术学院,成都611756

出　　处：《计算机系统应用》2020年第5期275-279,共5页Computer Systems & Applications

基　　金：国家重点研发计划(2017YFB1401401)。

摘　　要：在文本分类中,基于Word2Vec词向量的文本表示忽略了词语区分文本的能力,设计了一种用TF-IDF加权词向量的卷积神经网络(CNN)文本分类方法.新闻文本分类,一般只考虑正文,忽略标题的重要性,改进了TFIDF计算方法,兼顾了新闻标题和正文.实验表明,基于加权词向量和CNN的新闻文本分类方法比逻辑回归分类效果有较大提高,比不加权方法也有一定的提高.In the text classification methods,the text representation based on the Word2Vec ignores the weight of words in distinguishing text.The method of combining Word2Vec weighted by TF-IDF and CNN is designed.In news text classification,the importance of news title is always neglected.Therefore,this study proposes an improved TF-IDF method,which takes both news title and body into account.Experiments show that the news text classification method based on weighted word vector and CNN has a greater improvement than the logistic regression classification.And its effect increases by 2 or 3 percentage points than the un-weighted method.

关 键 词：文本分类 TF-IDF技术 Skip-gram模型 词向量 卷积神经网络 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术] TP18[自动化与计算机技术—计算机科学与技术]