本文已被：浏览 1631次 下载 2163次
Received:October 02, 2019 Revised:October 29, 2019
Received:October 02, 2019 Revised:October 29, 2019
中文摘要: 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TF-IDF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.
中文关键词: 文本分类 TF-IDF技术 Skip-gram模型 词向量 卷积神经网络
Abstract: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.
keywords: text classification TF-IDF Skip-gram word vector CNN
文章编号： 中图分类号： 文献标志码：
基金项目:国家重点研发计划（2017YFB1401401）
引用文本：
胡万亭,贾真.基于加权词向量和卷积神经网络的新闻文本分类.计算机系统应用,2020,29(5):275-279
HU Wan-Ting,JIA Zhen.News Text Classification Based on Weighted Word Vector and CNN.COMPUTER SYSTEMS APPLICATIONS,2020,29(5):275-279
胡万亭,贾真.基于加权词向量和卷积神经网络的新闻文本分类.计算机系统应用,2020,29(5):275-279
HU Wan-Ting,JIA Zhen.News Text Classification Based on Weighted Word Vector and CNN.COMPUTER SYSTEMS APPLICATIONS,2020,29(5):275-279
