基于文本摘要的新闻分析系统的研究与设计

s the main implementation of information transmission in modern human society, news communication has become an essential part of people\'s daily life. However, with the popularity and development of the Internet, online news websites and content pushing APPs are becoming widely popular,and the problem of information overload has become increasingly serious.This paper designs a news analysis system based on text summarization. The system can realize functions of automatic reports clustering and apply different summarizationapproaches for content analysis according to the clustering results. Moreover, to address the issues of incomplete content and redundant secondary information in the existing single-text summarization methods, this paper proposes anabstractive method based on graph convolution embeddingwhich effectively improves information selection ability, and confirmits effect on the CNN/Daily Mail datasets.