Hybrid top-N recommendation method based on social user tag
Yin Yan-teng · Jisuanji yingyong yanjiu · 2013
Aiming at the cold start user problem of tradition recommendation system,this paper proposed a hybrid top-N recommendation method which combined social user tag and collaborative filtering.The method obtained trustworthy user set through social relations.Based on user personalized tags,it applied a structural-context similar algorithm named SimRank to generate the similar neighbors set,which was used to generate predictive ratings.Finally,it combined the predictive ratings of traditional collaborative filtering to provide the recommendations.The experimental evaluations on all users and cold start users dataset demonstrate the proposed method outperforms the collaborative filtering approach in terms of recall and precision.