Personalization of information retrieval through user profiling

As the amount of data stored on electronic media increases, it becomes imperative to build information search and retrieval systems that provide user access to the right information quickly. When the search and retrieval process is based purely on key words that are devoid of context, it may result in missing potentially important sources or could result in finding too many unrelated sources. This article presents the NARAD distributed agent architecture that collaborates with users to identify their contextual framework for their information requirements. The architecture allows for individual agents to collaborate with each other to assist other agents based on profiles of users with knowledge of the topic in question. This paper describes the overall approach, the architecture, and results of the evaluation.