Individualized Automatic Classification of Web Documents

This paper applies Naive Bayes classifier in designing customized automatic web document classification to systematically collecting massive news articles from the Internet. The proposed news classification system allows users to establish the necessary information classifications based on their own preferences. When the amount of daily news is increasing, this approach enables users to effectively filter through large amount of articles and more focused on interested articles. Performances of the proposed approach are characterized by the recall rate and precision. This system can achieve over 66% recall rate, and over 89% precision rate for a real-world Chinese test database.