A scalable privacy-preserving recommendation scheme via bisecting k-means clustering

作者：

Highlights：

• A novel bisecting k-means clustering-based privacy-preserving CF scheme is proposed.

• A two-level preprocessing scheme is suggested to enhance scalability and accuracy.

• Effects of scalability and sparseness challenges are alleviated considerably.

• Accuracy of the solution is significantly better than knn-based CF and PPCF methods.

摘要

•A novel bisecting k-means clustering-based privacy-preserving CF scheme is proposed.•A two-level preprocessing scheme is suggested to enhance scalability and accuracy.•Effects of scalability and sparseness challenges are alleviated considerably.•Accuracy of the solution is significantly better than knn-based CF and PPCF methods.

论文关键词：Accuracy,Binary decision diagrams,Clustering methods,Data preprocessing,Data privacy,Recommender systems

论文评审过程：Received 13 December 2012, Revised 11 February 2013, Accepted 21 February 2013, Available online 27 March 2013.

论文官网地址：https://doi.org/10.1016/j.ipm.2013.02.004