A Novel Recommender Method in Collaborative Tagging Systems Based on Time Sensitive Topic Recommendation

摘要

Recently, collaborative tagging systems have grown in popularity on the web, on sites that allowing users to tag bookmarks, photographs and other contents. Algorithms based on tags have applied in many recommender systems, for tags represent both the contents of items and comprehension of users to items. Based on our experiments, we discover that users' interests are fall into many topics and change along with time variation. In this paper, we propose a new method based on topic recommendation, also we consider the time issue into incorporating with users' interests change. The dataset used in the experiments are extracted from the popular bookmark site Delicious, moreover, we adopt a novel measure to the method used in the research. The result demonstrated that the new algorithm is very effective.