Web mining for cyber monitoring and filtering

Like any self-regulating environment, the Internet is fertile ground for all kinds of potential abuse ranging from get-rich-quick scams, touting of illegal or adult-oriented material, promotion of extremist/anarchist views, to online pimping, etc. Consequently, the ability to discreetly intercept and analyze Internet access has tremendous potential in shielding users, especially our youngsters, from inappropriate content. This paper proposes one such system, the Web access monitoring and filtering (WAMF) system. The WAMF system comprises two main decoupled components, one for online monitoring and filtering, and the other for offline Web classification and data analysis. The former tracks, tallies, and selectively blocks user Web access in real-time, whereas the latter employs Web mining techniques to classify Web pages into pre-defined user categories and analyze user Web access data for user behavior patterns. In this paper, we will discuss the WAMF system, and in particular, Web mining techniques for adaptive Web page categorization.