News Recommender System Based on Topic Detection and Tracking

摘要

In web recommender systems, clustering is done offline to extract usage patterns and a successful recommendation highly depends on the quality of this clustering solution. As for collaborative recommendation, there are two ways to calculate the similarity for clique recommendation: Item-based Clustering Method and User-based Clustering Method. Researches have proved that item-based collaborative filtering is better than user-based collaborative filtering at precision and computation complexity. However, the common item-based clustering technologies could not quite suit for news recommender system, since the news events evolve fast and continuous. In this paper, we suggest using technologies of TDT to group news items instead of common item-based clustering technologies. Experimental results are examined that shows the usefulness of our approach.