Integrating Evaluation into Development of Knowledge-Based System for Adaptive E-learning System

Publication Type:
Issue Date:
Full metadata record
One of the main obstacles in knowledge-based systems, particularly in the field of evaluation, is to determine when the development of a knowledge base is complete with valid knowledge available to use. Ironically, this question is not easy to answer, since it involves many aspects in the lifecycle of knowledge-based systems. Some efforts have been introduced in literature, although none have been widely used or even agreed upon. In our investigation, we found that it is possible to provide a ready to use knowledge base without a separate testing phase. This thesis examines the way in which the testing phase could be integrated into the development phase through a statistical monitoring task during the knowledge acquisition process. This research tailored the dynamic evaluation framework and utilised the RDR technology as an incremental knowledge acquisition methodology, with additional statistical analysis to monitor the development of the knowledge base. Using this framework, this thesis presents INCAES, an incremental adaptive e-learning system. The knowledge base was developed incrementally for suggesting learning contents to students. The recommendation was provided by analysing the learner model as a representation of the students in adaptive learning. However, a systematic understanding of attributes of learners is still lacking in the field of adaptive learning, since those attributes are regularly changed and updated. Thus, we keep track by investigating the potential learner’s attributes that could be adapted to the adaptive system from a number of recent works. The evaluation of the theoretical concept has been conducted using two real world case studies. The first case study was applied on the Web Programming subject, while the second case study was applied on the Networking Essentials subject. The test cases generated from learners’ attributes and values were used to evaluate the knowledge base. Different scenarios were provided for the case study. The first scenario dealt with stopping the progress of the knowledge acquisition process at an early stage, while the second scenario dealt with stopping it at a later stage. Following the knowledge base construction, two steps of testing were conducted to measure the performance of the initial knowledge base and the subsequent test set. The results have shown that the framework allows for the rapid development of a knowledge base with valid knowledge inside.
Please use this identifier to cite or link to this item: