“I know what you need to buy”: context-aware multimedia-based recommendation system

摘要

Many people spend significant amounts of time navigating the worldwide web to see hypermedia, including online shopping. Multimedia shown in a web browser is one of the main sources of information for online shopping. In many cases, users are motivated to purchase a specific item by the impressions they receive when they are looking at related hypermedia: e.g. they may gain motivation to purchase a car after seeing a transportation item on a news website, especially when the users are looking at the images contained in the hypermedia. This leads to the motivation to build an intelligent system that is aware of what users are thinking when they are navigating hypermedia. This paper proposes an agent-based methodology that can first identify semantics of hypermedia images in order to derive contextual information, then recommends specific web services by anticipating needs or products in which the user may be interested. To identify and interpret the images, we propose extended ARG, which integrates ARG image content and semantics in the corresponding ontology. To show the feasibility of the idea, we developed CAMA (Context-Aware Multimedia Agent), a prototype agent-based system.