Research on Document Clustering from Internet Public Opinions

Generally, in the traditional multilingual topic discovery, it is the multilingual text for a single goal of the conversion and then clustering. On this basis, we have constructed a custom dictionary for the people with the highest percentage of Chinese, Japanese and English in this paper. At the same time, we have improved the single-pass clustering algorithm in single language. And considering the characteristics of news effectiveness, we have proposed a multilingual text composite clustering algorithm based on fusion time impact factor, which makes the clustering analysis results more reasonable, and better reflects the characteristics of the effectiveness of network news.