A contextual information retrieval service for educational environments

This paper presents the PortEdu Project, an Educational Portal, which consists of a MAS (MultiAgent System) architecture for leaning environments in the web, strongly based on personalized information retrieval. Experiencing search mechanisms it has been detected that the success of distance learning mediated by computer is linked to the contextual search tools quality. Our goal with this project is to aid the student in his learning process and retrieve info pertaining to the context of the problems studied by the students. Our experience about distance learning demonstrates that students with difficulties in specific problems solving, during the use of a distance learning environments, going though, in web research with the intention to find additional information on the studied topic. However, this search is not always satisfactory. The existing tools make a result classification not taking in consideration the specific needs of the user. Before this problem, it has been idealized a model to toil with these difficulties in the educational context. This model is based on two autonomous agents: User Profile Agent and Information Retrieving Agent. Such agents will be communicating with each other and among other agents of the learning environment though the multiagent FIPA-OS platform. In PortEdu, the search refinement is done automatically, based on available info in the users profiles, student's models (student cognitive level) and ontology (learning environment has its own ontology). So, the student makes a high-level information request and receives a distilled reply. A greet advantage on the project compared to other related work is the contextual and personalized content which is automatically retrieved, based on the application ontology, info on user profile agent, on the student model agent and on the attributed weight defined by the expert and the students for each useful document.