基于重要性的信任感知推荐方法  

Trust aware recommendation method based on significance

作　　者：谭晏松[1] 宋铁成 TAN Yan-song;SONG Tie-cheng(Artieficial Intelligence and Big Data College,Chongqing College of Electronic Engineering,Chongqing 401331,China;School of Communication and Information Engineering,Chongqing University of Posts and Telecommunications,Chongqing 400065,China)

机构地区：[1]重庆电子工程职业学院人工智能与大数据学院,重庆401331 [2]重庆邮电大学通信与信息工程学院,重庆400065

出　　处：《计算机工程与设计》2021年第8期2240-2247,共8页Computer Engineering and Design

基　　金：国家自然科学基金项目(61702065);重庆市教育委员会科技攻关基金项目(KJ1602906)。

摘　　要：针对基于信任模型的推荐系统认为所有项目对所有用户具有相同重要性的问题,提出一种基于重要性的信任感知推荐方法。基于项目重要性对所有用户不一样的假设提出了新的信任度量,利用相对于用户的人口统计上下文来测量项目对于用户的重要性,计算活动用户和每个集群的人口统计特征的相似性,将最相似集群中的用户视为候选邻居,根据活动用户的信任邻居的偏好为其生成一个推荐列表。实验结果表明,提出方法在大多数情况下都优于其它方法,具有较高的预测精度和质量。Aiming at the problem that all items are of the same importance to all users in the recommendation system based on trust model,a trust aware recommendation method based on significance was proposed.A new trust metric was proposed based on the assumption that item importance was different for all users.The significance of an item for a user was measured with respect to the demographic context of the user,and the similarity of the demographic characteristics of the active users and that of each cluster was calculated.Users in the most similar cluster were regarded as candidate neighbors.A recommendation list was generated for active users according to their preference of trusted neighbors.Experimental results show that the proposed method is superior to other methods in most cases and has high prediction accuracy and quality.

关 键 词：推荐系统 信任 协同过滤 人口统计特征 混合蛙跳算法 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]