Differentially Private Top-k Items Based on Least Mean Square——Take E-Commerce Platforms for Example

User preference data broadly collected from e-commerce platforms have benefits to improve the user’s experience of individual purchasing recommendation by data mining and analyzing, which may bring users the risk of privacy disclosure. In this paper, we explore the problem of differential private to...

Full description

Saved in:
Bibliographic Details
Published in:Wuhan University journal of natural sciences 2019-04, Vol.24 (2), p.98-106
Main Authors: Cao, Mengyue, Wu, Fusheng, Ni, Mingtao, Xiong, Xingxing, Liu, Shubo, Wang, Jun
Format: Article
Language:English
Subjects:
Algorithms
Biomedical and Life Sciences
Computer Science
Data mining
Data processing
Electronic commerce
Kalman filters
Life Sciences
Materials Science
Post-processing
Privacy
Recommender systems
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:User preference data broadly collected from e-commerce platforms have benefits to improve the user’s experience of individual purchasing recommendation by data mining and analyzing, which may bring users the risk of privacy disclosure. In this paper, we explore the problem of differential private top- k items based on least mean square. Specifically, we consider the balance between utility and privacy level of released data and improve the precision of top- k based on post-processing. We show that our algorithm can achieve differential privacy over streaming data collected and published periodically by server provider. We evaluate our algorithm with three real datasets, and the experimental results show that the precision of our method reaches 85% with strong privacy protection, which outperforms the Kalman filter-based existing methods.
ISSN:1007-1202
1993-4998
DOI:10.1007/s11859-019-1374-x