This paper presents a novel approach to web user profiling. Our proposed approach consists of two main parts. The first part focuses on discovering user interests in a user feedback collection, usually including relevant and irrelevant documents. Frequent pattern mining widely used in data mining community is applied to extract user feedback information. The second part is to represent user profiles. We introduce a novel user profile model based on belief function reasoning. In this model, the user profile is described by a probability distribution over the user feedback information extracted. Experimental results on an information filtering task show that the proposed approach clearly outperforms several baseline methods.
CITATION STYLE
Pipanmaekaporn, L., & Kamonsantiroj, S. (2015). A belief function reasoning approach to web user profiling. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9375 LNCS, pp. 509–516). Springer Verlag. https://doi.org/10.1007/978-3-319-24834-9_59
Mendeley helps you to discover research relevant for your work.