A web-page recommender system via a data mining framework and the Semantic Web concept

Traditional search engines require users to form the keyword based query which can accurately depict the search topic. More importantly, search engines are generally unable to customise the results according to the users' preferences. Recently, an alternative approach of retrieving the information, known as the recommender system is proposed. A recommender system is an intermediary program that intelligently generates a list of information which matches the users' preferences. In this paper, a new recommender system framework based on data mining techniques and the Semantic Web concept is proposed. Two information filtering methods for providing the recommended information (i.e., content-based and collaborative filtering) are considered. Both filtering techniques are based on data mining algorithms which provide efficiency in handling large data sets. In addition, the Semantic Web concept, in which the information is given well-defined meaning, is incorporated into the framework in order to provide the users with semantically-enhanced information. To demonstrate the potential use of the proposed framework, a system prototype for recommending the University of Miami's Web pages was implemented to enhance the performance of the traditional query-based information retrieval approach provided on the website.