Information Filtering Using the Dynamics of the User Profile

摘要

This paper presents an adaptive algorithm for learning the user profile. The user profile is learned incrementally and continuously based on user's initial profile, his actions and on semantic interpretation of queries using hypernyms extracted by WordNet. A novel model, time - words vector hyperspace, is introduced in order to keep track of the user's interests changes. This new model is achieved by adding a temporal dimension to the classical vector hyperspace model. The results of the retrieval experiments using this new algorithm show an improved effectiveness over the current information retrieval techniques.