An Enhanced Collaborative Filtering Algorithm Based on Time Weight


Because traditional collaborative filtering algorithm does not consider the influence of time on user’s interests, taking equal consideration with different user’s interests in different time leads to the neighbors may not be the similar neighbors set. For solving the problem, the enhanced algorithm endows each score with a time weight which declines gradually as time going, and uses the weighted score to search the nearest neighbors. Experiments of Movielens dataset show that the enhanced algorithm added time weight improves quality of recommendation system.

1 Figure or Table

Cite this paper

@article{HuaiZhen2009AnEC, title={An Enhanced Collaborative Filtering Algorithm Based on Time Weight}, author={Yang Huai-Zhen and Li Lei}, journal={2009 International Symposium on Information Engineering and Electronic Commerce}, year={2009}, pages={262-265} }