Record:   Prev Next
Author Abdel-Razek, Mohammed
Title Multi-agent approach towards intelligent e-learning system
book jacket
Descript 201 p
Note Source: Dissertation Abstracts International, Volume: 65-10, Section: B, page: 5224
Adviser: Claude Frasson
Thesis (Ph.D.)--Universite de Montreal (Canada), 2004
Web technology is becoming more universal in the world; accordingly, the use of the Web to provide learners and helpers (tutors or others) with a real-time learning environment at a distance is also continuing to grow rapidly. Furthermore, the Web as a distance learning environment has wonderful potential to replace traditional means of teaching distance learners. However, the ability to offer collaboration along with adaptation within the same system is still a challenging and important problem. To address this issue, we developed a new approach called dominant meaning. This research has applied a multi-agent technique based on the dominant meaning approach to design a multi-user e-learning system to provide learners with a Collaboration Adaptive Distance Learning Environment (CADLE). The proposed system is intended to create more collaboration between online learners and an individualized approach to adapt course presentation without the need for additional input from a user
This dissertation consists of three areas of research regarding the implications of CADLE
The first area establishes flexible domain knowledge and a dynamic user profile in order to establish a cohesive collaboration and provide individual adaptation. This is achieved by a combination of the dominant meaning approach and machine learning algorithms. This combination is a novel approach in the field of machine learning
Second, based on the dominant meaning approach, this research has investigated the technical feasibility and also the utility of applying a multi-agent technique to perform CADLE. An e-learning system called Confidence Intelligent Tutoring System (CITS) has been implemented to determine the viability of this approach. The CITS has been implemented, tested, and evaluated. The prototype of CITS is based on six types of agents: user interface agent, cognitive agent, behavior agent, guide agent, context-based information agent, and confidence agent
Third, results from an experiment conducted on each agent are presented separately. Moreover, the results of the evaluation of the whole system show that the CITS is a functional and robust system. In the same sense, the utility of using dominant meaning approach along with machine learning algorithms to improve Web search results and to analyze users' browsing activity in CITS was proven, and the technical feasibility has been established
School code: 0992
Host Item Dissertation Abstracts International 65-10B
Subject Computer Science
Alt Author Universite de Montreal (Canada)
Record:   Prev Next