Individual privacy and organizational privacy in business analytics

Abstract

The use of business analytics to reveal useful information from transactional and operational databases involves accessing private information of both individuals and organizations. This practice has raised the privacy concern of individuals and policy makers. In this paper, we propose a privacy preserving approach for business analytics based on keyed Bloom filters. We illustrate this approach using market basket analysis, and then evaluate its performance using real datasets from a point-of- sale and web clickstream. The major advantage of this approach is the positive relationship for the level of privacy security and analysis precision. In other words, the tradeoff of the level of privacy protection for our solution is data storage space; whereas for other existing methods, there is always a tradeoff between analysis precision and the level of privacy protection.