Popularity-Based Summarization of Chinese Text: Implicit Weight-Based Features for Newspaper Articles

The fully idiographic nature of Chinese language facilitates the algorithms in natural language processing by a tight correspondence between characters and text meaning, and also allows for substantially less preprocessing as compared to the case of other languages. We have recently developed and tested a popularity-based concept to summarization of Chinese text. This paper argues that the popularity based concept, if applied to news articles, naturally combines with the weight-based summarization techniques, and yields an excellent efficiency in Chinese text summarization. The summarization program is available online as java applet (http://cpu.icu.ac.jp/cn/).