No abstract

Hyperlink recommendation overcomes the problem of quick and easy access to information in web systems. A method that integrates web usage and content mining was proposed and examined in this paper. Potentially interesting documents are prompted to the user on the basis of usage patterns and conceptual spaces matched against the active user session. Automatic term selections and web usage distinction according to the time of visit were introduced to enhance method effectiveness.

No abstract

No abstract

In this paper a fuzzy-based recommendation method is presented. Its main goal is to improve the recommendation recall maintaining high recommendation precision. The formal model has been built to describe the method and to analyze how the measures used in traditional Information Retrieval may be adapted to evaluate the effectiveness of recommendation process. The original contributions consist among others of proving several properties which show that the method is able to adapt to changing user's needs and achieving the maximum effectiveness if the component methods work properly.