Personal Document Recommendation System Based on Data Mining Techniques

Most existing recommendation systems may not be very effective due to the lack of the adequate knowledge of user´s behavior or interests. Some systems are not] efficient enough because of the huge on-line computational demand. In this paper, we propose a personal documents recommendation system that effectively filters the on-line news on WWW for each individual user. The proposed system recommends useful news by employing the profiling techniques of modified content- and collaborative-based filtering. In order to reduce the on-line computation and improve the recommendation quality, we design a tree-based data mining algorithm that treats users´ behavior and interest as input and filter the news documents efficiently.