A Survey of Privacy Preserving Utility Mining

Abstract

High-utility pattern mining has emerged as an important research topic in data mining. It aims at discovering patterns having a high utility (e.g. profit or weight) in transaction or sequence databases. HUPM can be applied in various fields such as market basket analysis, website clickstream analysis, stock market analysis, retail and bioinformatics. In the era of information technology, it has become easy to locate and access information. A greater access to information has many benefits. However, it may also lead to privacy threats if datasets containing sensitive and important information are shared and made public. Therefore, privacy preservation has become a critical challenge for data mining. This chapter provides an up-to-date survey on privacy preserving utility mining (PPUM). The main purpose is to provide a general overview of recent techniques and algorithms for PPUM. The chapter focuses on research on both privacy preserving high-utility itemset mining and privacy preserving high-utility sequential pattern mining. Key concepts and terminology are introduced and discussed. Moreover, latest solutions for PPUM are compared. Finally, challenges and opportunities related to PPUM are discussed.