A Comparative Study on Automatic Categorization Methods for Chinese Search Engine

摘要

To improve the service of A Chinese search engine we developed and analyze the thousands of web pages it collected, we implemented and evaluated four wide-used text classification algorithms, i.e. SVM, K-NN, Naive Bayes, and Simple TFIDF. Experimental results show that Simple TFIDF is more suitable to our task than others, because it takes the advantages of lower running time and higher generalizing capacity.