An intelligent subtopics suggester for an e-learning system

Abstract

ELearning, is a planned teaching or learning experience that uses a wide spectrum of computerbased technologies to reach learners. The focus of this project was the case where students are logging into and using the system independently of other students and staff members. This fits firmly into the general definition of the asynchronous e-learning environment, where the instructor develops content and the students access it online. The instructor also suggests content to students with specific needs, to do this, current systems use face-to-face discussions, online discussions or emails. Because of the increased number of students involved in eLearning this individualized attention to students’ needs has become untenable and unsustainable hence the need for an Intelligent Subtopics Suggester. The model lacks, from a literature point of view. This research aimed at addressing this gap through design of an Intelligent Subtopics Suggester model. This project developed an Intelligent Subtopics Suggester model that is appended to an eLearning system and that analyzes the user's questions and suggests help subtopics to the lecturer/tutor. The intelligent model was based on a knowledge base that was made using knowledge from domain experts. The knowledge base captured the keywords and terminologies that describe a subtopic of interest. These keywords and terminologies were matched against the keywords in the students' question. This helped to identify the students’ key learning needs, by identifying the subtopics learners have not understood well, based on the frequency that a question is asked from a given subtopic. Based on the above model a prototype was developed, the prototype has 6 production rules for inference and 1505 facts (keywords) in its knowledge base. The Intelligent Subtopic Suggester was able to identify the topics from which questions had been asked. To evaluate the prototype three accuracy standards were applied, these were precision, recall and accuracy, on average the Intelligent Subtopic Suggester has a precision of 0.727, a recall of 0.972 and an accuracy of 0.766. The Intelligent Subtopic Suggester was found to have acceptable values of precision and accuracy and hence can be considered reliable. The findings of this research will be of great importance to eLearning system developers and the research community. Keywords: Intelligent Subtopic Suggester, eLearning, Intelligent tutor systems, Artificial intelligence, web based learning, precision, recall, accuracy