A State-of-the-Art Review of Data Stream Anonymization Schemes

A State-of-the-Art Review of Data Stream Anonymization Schemes

Aderonke B. Sakpere (University of Cape Town, South Africa) and Anne V. D. M. Kayem (University of Cape Town, South Africa)
Copyright: © 2014 |Pages: 27
DOI: 10.4018/978-1-4666-6158-5.ch003
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

Streaming data emerges from different electronic sources and needs to be processed in real time with minimal delay. Data streams can generate hidden and useful knowledge patterns when mined and analyzed. In spite of these benefits, the issue of privacy needs to be addressed before streaming data is released for mining and analysis purposes. In order to address data privacy concerns, several techniques have emerged. K-anonymity has received considerable attention over other privacy preserving techniques because of its simplicity and efficiency in protecting data. Yet, k-anonymity cannot be directly applied on continuous data (data streams) because of its transient nature. In this chapter, the authors discuss the challenges faced by k-anonymity algorithms in enforcing privacy on data streams and review existing privacy techniques for handling data streams.
Chapter Preview
Top

Introduction

The need for data protection, especially when needed for analysis, research and data mining purposes has led to the development of several privacy enforcing schemes. Considerable attention has been given to static data protection (Issa, 2009; Iyengar, 2002; Samarati, 2001; Sweeney, 2001, 2002a, 2002b). Static data are non-real time and so the constraints for processing and/or analysis are not time sensitive. Conversely, there is a lot of data that evolves with time and space, typically referred to as data streams with many real world applications (Guo & Zhang, 2013).

Data streams are real-time and continuous data flows that are ordered implicitly by arrival time or explicitly by timestamps (Golab & Özsu, 2003). The order in which streaming data arrives cannot be pre-determined (Golab & Özsu, 2003). Streaming data emerges from various electronic sources (such as mobile phones or computers) and is expected to be processed online in real-time with minimum delay (Zakerzadeh & Osborn, 2013). In Figure 1, we illustrate that streaming data emerge from a source and essentially has a target destination. Examples of applications that use data streaming include web applications, financial applications and security applications (Zakerzadeh & Osborn, 2013). Data streams can also be a form of temporal data (Wang, Xu, Wong, & Fu, 2010). Temporal data is time-critical because the snapshot available at each timestamp must be made available for necessary action (Wang et al., 2010).

Figure 1.

Illustration of data streams

Complete Chapter List

Search this Book:
Reset