山东农业大学学报（自然科学版）2017，Vol.48Issue(2)：192-198,7.DOI:10.3969/j.issn.1000-2324.2017.02.006
概率粗糙集模型在推荐算法中的应用
Application of Probabilistic Rough Set Model in Recommendation Algorithm
摘要
Abstract
Recommendation algorithm can dig into the potential interest of users and then automatically recommends the project to the users.It is one of the intellectual approaches solving the information overload.As the number of users and number of items are more,the sparseness of rating matrix has a strong impact on the effect of recommendation.The priori knowledge of recommendation is missing seriously.Rough set is an effective method which can adopt an incomplete knowledge to carry out ratiocination.It proposes dividing the boundary region by using probabilistic rough set threshold α and β,creating recommendation strategy and reducing the effect of the sparseness of score matrix to the recommendation result.The experimental results indicated that the model of probabilistic rough set could improve the recommendation accuracy rate under the circumstance of the high sparseness of score matrix.The recommendation accuracy rate could be up to 92.80％ in the Movie Lens data sets and the highest fraction of coverage could be up to 100％.关键词
概率粗糙集/推荐算法/参数学习Key words
Probabilistic Rough Set/recommendation algorithm/parameter learning分类
信息技术与安全科学引用本文复制引用
陈功平,王红..概率粗糙集模型在推荐算法中的应用[J].山东农业大学学报（自然科学版）,2017,48(2):192-198,7.基金项目
2015年度安徽高校自然科学研究重点项目(KJ2015A435) （KJ2015A435）
安徽省2016年高校优秀青年人才支持计划重点项目(gxyqZD2016570) （gxyqZD2016570）
安徽省2014年高校优秀青年人才支持计划 （）
