A novel recommendation system with collective intelligence

Academic resources on web include courses, educational videos, scientific literatures, experts, peers and all of the useful stuff for research. It is crucial for researchers especially freshman to access and control the academic resources when they start and conduct a research subject. This paper proposes a recommendation system suggests high-quality materials to users according to their research interest. The system makes use of an ontology which is created by domain experts to define the categories of the entire research subjects. The materials of each category are recommended by domain experts and users which are called collective intelligence. And the recommended academic resources list (RARL) is updated adaptively with the operation of the system. Based on the results of user intention detection the system assign each user to the corresponding categories and the user gets his recommendation based on the content in RARL. In the proposed system there are 15,000 education videos, 123GB related materials of 917 courses and 30,000 web pages. The experimental tests show that the system performance is well: the performance is not getting worse when there are more web pages.