Preprint Article Version 1 Preserved in Portico This version is not peer-reviewed

Item-Oriented Personalized LDP for Discrete Distribution Estimation

Version 1 : Received: 4 June 2023 / Approved: 5 June 2023 / Online: 5 June 2023 (09:36:31 CEST)
Version 2 : Received: 5 June 2023 / Approved: 8 June 2023 / Online: 8 June 2023 (03:16:47 CEST)

How to cite: Li, X.; Zhu, H.; Zhang, Z.; Xie, M. Item-Oriented Personalized LDP for Discrete Distribution Estimation. Preprints 2023, 2023060289. https://doi.org/10.20944/preprints202306.0289.v1 Li, X.; Zhu, H.; Zhang, Z.; Xie, M. Item-Oriented Personalized LDP for Discrete Distribution Estimation. Preprints 2023, 2023060289. https://doi.org/10.20944/preprints202306.0289.v1

Abstract

Discrete distribution estimation is a fundamental statistical tool, which is widely used to perform data analysis tasks in various applications involving sensitive personal information. Due to privacy concerns, individuals may not always provide their raw information, which leads to unpredictable biases in the final results of estimated distribution. Local Differential Privacy (LDP) is an advanced technique for privacy protection of discrete distribution estimation. Currently, typical LDP mechanisms provide same protection for all items in the domain, which imposes unnecessary perturbation on less sensitive items and thus degrades the utility of final results. Although, several recent works try to alleviate this problem, the utility can be further improved. In this paper, we propose a novel notion called Item-Oriented Personalized LDP (IPLDP), which independently perturbs different items with different privacy budgets to achieve personalized privacy protection. Furthermore, to satisfy IPLDP, we propose the Item-Oriented Personalized Randomized Response (IPRR) based on the observation that the sensitivity of data shows an inverse relationship with the population size of respective individuals. Theoretical analysis and experimental results demonstrate that our method can provide fine-grained privacy protection and improve data utility simultaneously.

Keywords

Discrete distribution estimation, Local differential privacy, Item-oriented personalization, Randomized response

Subject

Computer Science and Mathematics, Other

Comments (0)

We encourage comments and feedback from a broad range of readers. See criteria for comments and our Diversity statement.

Leave a public comment
Send a private comment to the author(s)
* All users must log in before leaving a comment
Views 0
Downloads 0
Comments 0
Metrics 0


×
Alerts
Notify me about updates to this article or when a peer-reviewed version is published.
We use cookies on our website to ensure you get the best experience.
Read more about our cookies here.