基于用户相似性的加权SlopeOne算法-计算机应用研究.PDFVIP

优先出版 计 算 机 应 用 研 究 第33 卷 基于用户相似性的加权Slope One 算法* 1 1, 2 2 刘林静 ，楼文高 ，冯国珍 (1．上海理工大学 光电信息与计算机工程学院，上海 200093 ；2 ．上海商学院 管理学院，上海，200235) 摘 要：针对因Slope One 算法没有考虑相似性，而导致个性化推荐准确率不高的问题，提出了一种基于用户相似性的 加权Slope One 算法（BUS Weighted Slope One 算法），通过先评定用户活跃度，筛选出活跃用户，然后依据项目间相似 性对部分未评分项目进行预测填充，再利用用户间的相似性得到用户的最近邻居集合，将用户间的相似性作为预测评分 权重，最后根据最近邻居集合对样本进行预测分析。通过三组实验研究结果表明，与其他传统方法相比，本文提出的方 法同时提高了评分预测的准确性和计算效率。 关键词：个性化推荐系统；Slope One ；相似性；用户活跃度；评分预测 中图分类号：TP391 New weighted Slope One algorithm based on user similarity 1 1, 2 2 Liu Linjing , Lou Wengao , Fen Guozhen (1. Photoelectric Information Engineering College with Computer, Shanghai University of Science Technology, Shanghai 200093, China; 2. Faculty of Management, Shanghai Business School, Shanghai 200235, China) Abstract: In order to solve the defect of the Slope One algorithm that the effect of recommending is not well in personalized recommendation system because of without considering the similarity. this paper proposes a Weighted Slope One algorithm based on the similarity of users (BUS Weighted Slope One algorithm).First assess the activity of users and pick out the active user. and then fill the matrix basing on the similarity of items. and then get the nearest neighbors according to the similarity of users and put the similarity of users as the weight of predicting rating. and lastly use samples to predict rating. Experimental results show that the proposed approach in this paper not only ensures the efficiency but also increases the rating predict