计算机工程与应用 ›› 2018, Vol. 54 ›› Issue (8): 214-219.DOI: 10.3778/j.issn.1002-8331.1609-0445

• 工程与应用 • 上一篇    下一篇

基于用户相似度和信任度的药品推荐算法

肖晓丽1,2,周锡玲1,2   

  1. 1.综合交通运输大数据智能处理湖南省重点实验室(长沙理工大学),长沙 410114
    2.长沙理工大学 计算机与通信工程学院,长沙 410114
  • 出版日期:2018-04-15 发布日期:2018-05-02

Medicine recommendation algorithm based on user similarity and trust

XIAO Xiaoli1,2, ZHOU Xiling1,2   

  1. 1.Hunan Provincial Key Laboratory of Intelligent Processing of Big Data on Transportation(Changsha University of Science and Technology), Changsha 410114, China
    2.College of Computer and Communication Engineering, Changsha University of Science and Technology, Changsha 410114, China
  • Online:2018-04-15 Published:2018-05-02

摘要: 为了解决协同过滤算法推荐精度低的问题,提出基于用户相似度和信任度的药品推荐算法。该方法通过离线使用DBSCAN算法对药品进行聚类来降低时间复杂度。引入共同评分药品阈值使用户相似度计算更准确,同时设置相似度阈值来限定相似性邻居的选取以克服KNN算法选取邻居的缺陷。根据用户的推荐可信度和评分可信度建立信任计算模型,计算基于相似邻居集的可信邻居集。通过两次邻居选择策略为目标用户产生药品推荐。仿真结果表明,该算法与其他算法相比在平均绝对误差、准确率和召回率上有更好的性能,提高了系统推荐精度。

关键词: 协同过滤, 信任计算模型, 用户相似度, 药品推荐

Abstract: Aiming at the problem of lower recommendation precision of collaborative filtering, medicine recommendation algorithm based on user similarity and trust is proposed. The method clusters drugs into several groups by using DBSCAN algorithm offline to reduce the time complexity. For the sake of computing the user similarity more precisely, co-rated drugs threshold is introduced to build similar neighbor set of target user, at the same time similarity threshold is introduced to restrict the selection of similar neighbors, which overcomes the defects of KNN algorithm. Then the trust computing model is designed according to recommendation credibility and score reliability. The trustworthy neighbor set of target user is selected in accordance with the degree of trust between users. Finally, drugs are recommended to target user through twice neighbor selection strategy. Experimental results show that compared with the existing algorithms, the proposed algorithm has better performance in mean absolute error, precision and recall ratio, which improves the recommendation precision.

Key words: collaborative filtering, trust computing model, user similarity, medicine recommendation