Study on the Design and Implementation of a New Anonymization Algorithm for Semi-Structured Transaction Data
Sung Hyun Hong, Gyu Sung Lee, D. Kim, Yong Wan Joo, Soon Seok Kim · Asia-pacific Journal of Convergent Research Interchange · 2023
The transaction data refers to information stored in a relational database, considering various customers' purchases at a supermarket, where multiple items are grouped together as a kind of set within a single cell.This paper aims to address the issue of anonymization in such transaction data.Y.He and J. Naughton have previously proposed a technique known as 'Local Generalization,' which is based on k-anonymity to tackle this problem.However, this technique offers strong privacy protection and short execution times but suffers from significant information loss.On the other hand, J. Liu and K. Wang have proposed the HgHs (Heuristic generalization with Heuristic suppression) algorithm to mitigate these drawbacks, with reduced information loss but longer execution times.In order to improve these existing shortcomings, we have proposed a new technique that can guarantee strong privacy while minimizing information loss compared to the above two techniques.In this paper, we will reconfigure the system so that it can be commercialized and present the design, implementation process, and test results of the existing proposed algorithm.The test used the datasets (BMS-WebView2 and BMS-POS) used in the existing local generalization and HgHs techniques, and the test results confirmed the accuracy of the proposed algorithm.