IntelliShopper
Filippo Menczer, W. Nick Street, Narayan Vishwakarma, Alvaro E. Monge, Markus Jakobsson · 2002
The IntelliShopper is a shopping assistant designed to empower consumers. It is a personal assistant in that it observes the users while shopping and learns their preferences with respect to various features that characterize shopping items. It is proactive in that it remembers the users' requests and autonomously monitors vendor sites for new items that might match the users' needs and preferences. Finally, it protects users' privacy by means of pseudonymity, IP anony\-mizing, and trusted filtering. Pseudonymity is achieved through the use of personae; we show that this approach also behooves successful classification. IP anonymizing can be performed in at least two manners, which we discuss and compare in the context of our application. Trusted filtering --- as opposed to merchant-based filtering --- improves privacy by allowing users to select their preferred privacy representative. This paper introduces the IntelliShopper system, discusses its architecture and components, describes a prototype implementation, and outlines preliminary evaluations of its performance.