Abstract—In this paper, we propose a web recommendation
system where user navigational patterns can be extracted from
web logs. First, the recommendation system discovers user
concepts from web logs step-by-step, and then extracts the
navigation patterns among these concepts. These navigational
patterns are then used to generate recommendation web pages
by matching the navigation behavior of a user personal
knowledge base. The pages in a recommendation list are ranked
according to their hub scores which are computed based on page
connectivity information. The experimental results show that
the web pages recommended by our system are of better quality
and acceptable for humans from various domains, based on
human evaluators ranking as well as quality-value-based
Index Terms—Information retrieval, recommendation system, web mining, web search.
Yin-Fu Huang is with National Yunlin University of Science and Technology, Yunlin, Taiwan 640 (e-mail: firstname.lastname@example.org).
Jia-Tang Jhang is with Tornado Technologies Co., Ltd, Taipei, Taiwan 106 (e-mail: email@example.com).
Cite: Yin-Fu Huang and Jia-Tang Jhang, "Recommendations of Personal Web Pages Based on User Navigational Patterns," International Journal of Machine Learning and Computing vol.4, no. 4, pp. 307-312, 2014.