文献类型：期刊文献

英文题名：Improved Collaborative Filtering Algorithm Incorporating User Information and Using Differential Privacy

作者：Ren, Jiahui[1]; Xu, Xian[1]; Yu, Huiqun[1]

机构：[1] Department of Computer Science and Engineering, East China University of Science and Technology, Shanghai, 200237, China

年份：2019

卷号：1042 CCIS

起止页码：458

外文期刊名：Communications in Computer and Information Science

收录：EI(收录号：20195007831605)

语种：英文

外文关键词：Information use - Recommender systems - Signal filtering and prediction

摘要：Collaborative filtering algorithm is one of the most popular recommendation algorithms. There is, however, the risk of privacy leakage when making effective recommendation. Differential privacy is a relatively new privacy protection mechanism in the field, and has been used in recommendation systems. To this end, the existing research still has some disadvantages. Particularly, they do not have satisfactory performance and have difficulty in solving the cold start scenarios. In this paper, we propose an improved differential privacy enabled collaborative filtering algorithm incorporating user information. The algorithm improves similarity calculation, and solves the user cold start problem by making effective use of user information (related attributes). Experiments show that with the same privacy guarantee, the proposed algorithm improves the performance of the recommendation system and indeed solves the problem of cold start to some good extent. ? 2019, Springer Nature Singapore Pte Ltd.