摘要
针对传统协同过滤算法稀疏矩阵和推荐精度不高的问题,根据一种社会心理学模型提出了基于群体动力学的协同过滤算法。该算法综合考虑了个体因素和环境因素对用户评分行为的影响,以此来调整传统的评分预测方法,然后为用户进行推荐。实验结果表明,该算法可以明显地提高推荐的精确度,有效地缓解稀疏矩阵问题;同时该算法还可以有效减少积累误差。最后将该算法成功运用在西安景点的推荐服务上。
As to the issue of scarce matrix and low recommendation accuracy in traditional collaborative filtering algorithm,this paper came up with an collaborative filtering algorithm based on group dynamics of social psychology model. The algorithm adjusted the traditional prediction methods taking fully individual factors and environmental factors into consideration,then recommended to users. Experimental results show that this algorithm can not only achieve better recommendation accuracy and alleviat the data sparsity obviously,but also reduce error accumulation effectively. At last,the algorithm was successfully applied to recommend service of Xi’an attractions.
出处
《计算机应用研究》
CSCD
北大核心
2014年第12期3603-3605,3612,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(41271387)
西安市科技计划基金资助项目(SF1228-3)
陕西师范大学院士创新基金资助项目(999521)
关键词
K-近邻
协同过滤
群体动力学
推荐系统
K-nearest neighbor
collaborative filtering
group dynamics
recommendation system