Studies on Chinese Web page classification
Dou Shen, Yan Cong, Jian-Tong Sun, Yuchang Lu · 2004
In this paper we make studies on several key aspects for Chinese Web page classification such as Web page representation, word segmentation and feature selection. For the first two aspects, we test the published techniques on these issues on our Chinese corpora and give reasonable analysis for their performance. As to feature selection, we bring forward the idea of taking the role of a word's POS into consideration in pre-processing and the experimental results validate our idea.