Classification of characteristic words of electronic newspaperbased on the directed relation

摘要

Newspaper articles are gradually opened to Web system. But thecategories of articles are remained in conventional category types. Sousers of these system have to consider the appropriate searchingkeywords and categories when they search some articles. In this paper,we propose the new classification method for the characteristic wordsbased on directed relation. We think this classification can changecategories of articles and accessing keywords of users. First, weintroduce the word vector and directed relation degree. Second, wedefine the classification method, reference degree and abstractionmethod of cluster. Finally, we experimented using one month data and canget the following results. (1)Both directional tightly connected wordsare suitable base words of clusters of articles. (2)Level-0 abstractionmethod can draw the rough border of clusters. (3)One directional tightlyconnected words represent the main general words or proper words