A User-Adaptive Self-Proclamative Multi-Agent Based Recommendation System Design for E-Learning Digital Libraries
Ramalingam Ponnusamy, T. Venu Gopal · 2006
E-learning, the most modern Web application for dissemination of information caters to the needs of different learning communities. Web digital library is an instrument that provides various types of information required for e-learning. It is necessary to formulate a new technique, by taking into account various user communities and their personal requirements. This paper describes a new technique for design and development of a user-interface system called "a user adaptive self-proclamative multi-agent based recommendation system design" for e-learning digital libraries. This system is mainly designed for research people to collect the research literature. Basically this is a multi-agent based recommendation system that provides necessary suggestions by learning user-personal profiles and actions. The user has the provision to search the information and while searching, the system recommends the list of research literature that the user is interested. The profiles are concept patterned and a concept matrix is designed to represent each user as well as the subject. An ACM Computational Review Classification hierarchy along with conceptual matrix is used to represent the agent internal semantic model and to support the best inference. It is based on formulating a concept matrix, which consists of technical phrases and indexes occurring in the title and abstract of the paper. These phrases and indexes are picked up from the ACM computing review classification index, keywords as well as using phrases of Microsoft on-line computer dictionary. The columns of the matrix represent phrases and the rows represent the frequency of occurrences of the specified phrases in the document or query. The system is also doing concept relativity analysis. The usage of concept matrix enables the system to recommend/retrieve the exact information. It does a behavior prediction for various types of users and builds the user profile accordingly. It builds a global as well as personal profile. It also understands the user assistance requirements in various circumstances and adapts the user accordingly