An agent architecture for personalized Web stores
Generate an AI Snapshot to get a quick, structured summary of this paper.
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
This system exploits user modeling and flexible hypermedia techniques to tailor to the user the suggestion of goods and the description of the store catalog and argues that only an architecture composed of specialized agents can carry on personalized interact.
Abstract
We describe the architecture of a conl$urable Web store support.ing personalized interact.ions with users. Our system exploits user modeling and flexible hypermedia techniques to tailor to the user the suggestion of goods and the description of the store catalog. Customizing the system's behavior requires the parallel execution of several complex tasks during the interaction (e.g. identifying the user's preferences and dynamically generating the hypertextual pages of the store catalog). We argue that only an architecture composed of specialized ngrnts mn succcssfillly carry on t.hese tasks. Our system is developed in a Java-based environment using tools for building interoperable, agent-based systems (e.g. JDBC drivers for database independence; Voyager for the multiagent architecture).
