SEARCHY: An Agent to Personalize Search Results

User's behaviour in browsing sessions is a valuable source of information useful to analyze user interests and personalize the human-computer interaction during information seeking tasks. In this paper we present a simple way to machine-learn user profile and employ it to improve Web searches. Searchy is a personal agent, embedded in the Web browser, able to sense user's tastes, to assist her/him during query formulation phase and, finally, to filter and sort the results according to the user's needs. Searchy uses the DART P2P network to store user's profiles in order to enable collaborative recommendations.