earticle

영어

Critical surveys show that mobile agents are unable to delegate properly the user’s requests resulting into having a semi-automated system requiring the intervention of the clients. This paper describes a new brokering architecture called “Meta Broker”. It interacts with relevant online real estate websites, based on the user’s desires to recommend possible matching advertisements. The proposed architecture includes multiple agents: Shopping Bots; Product Brokering; Recommender System and Data Mining. The objective is to present an intelligent gateway that can elicit the client desires and finds highest possible matches for advertisements on the real estate market. The Meta Broker utilizes collaborative filtering techniques on previously achieved data as well as upon previous client feedbacks to categorize the new client behavior. The extracted knowledge directly affects the ranking of the adverts that are compatible with the user’s characteristics and the client queries.