Author(s)

L. Peyton & J. Hu

Abstract

There are now many services for individuals and businesses available over the Internet. In providing these services, personal data is exchanged. There are security and privacy concerns about protecting identity and controlling the use of personal data. The Liberty Alliance project has developed a set of standards and architecture for federated identity management. A circle of trust (CoT) is a network based on the Liberty Alliance architecture in which businesses collaborate to provide services in a manner that protects identity and carefully controls the sharing of personal data. We analyse the requirements of each stage of a knowledge discovery process in the context of a CoT, and recognize data collection as an important step with respect to protecting privacy and identity. An eHealth scenario in which data mining is used to detect prescription misuse illustrates how to implement a trusted data collection architecture in a CoT . Keywords: circle of trust, data collection, privacy, data mining, architecture. 1 Introduction With the growth of on-line services data is often shared and collected across organizations in distributed business to business (B2B) networks across the Internet. Data mining is an important tool for monitoring the services provided within such networks, but the data needed for data mining may come from several organizations. Because each organization operates independently, they each have only a partial view of the data involved. Data sharing is required in order to collect and analyse data. However, there are security and privacy concerns about protecting identity and controlling the use of personal data. In response to this concern, governments have enacted privacy laws such as the European Union Prime Directive on Privacy [1], the Health Insurance Portability and Privacy Act [2] in the United States and the Privacy Act and the Personal Information Protection and Electronic Documents Act (PIPEDA) [3] in Canada.

Keywords

circle of trust, data collection, privacy, data mining, architecture.