A top-N hot items recommendation algorithm based on items' attraction and popularity

作者：Gao Quanli; Gao Ling^*; Yang Jianfeng; Wang Hang

来源：Journal of Computational Information Systems, 2014, 10(11): 4913-4920.

DOI：10.12733/jcis10680

• Cold start problems
• Hot items
• Normalized hot value
• Normalized values
• Personalized recommendation
• Quality of recommendations
• Recommendation algorithms
• Recommendation techniques

Recommender systems apply data analysis and behavior analysis to make personalized recommendation for users, and collaborative filtering is one of the most popular recommendation techniques which have been widely used. However it has been suffered from data sparsity and cold-start problems which lead to low novelty and predict precision. To solve those problems, a new hot items recommendation algorithm based on the items' attraction and normalized values of hot which represent the items' popularity is proposed. The experimental results in real dataset show that the proposed algorithm can achieve better recommendation variety and reduce the average popularity. Furthermore, it can bring better quality of recommendation.

• 出版日期2014
• 单位西北大学