Roulette Wheel Selection Algorithm (RWSA) and Reinforcement Learning (RL) for personalizing and improving e-learning system

Ballera, Melvin Abes
Journal Title
Journal ISSN
Volume Title
Strathmore University
Various mechanisms to improve the learning process with the main objective of maximizing learning and dynamically selecting the best teaching operation to achieve learning goals have been done in the field of personalized learning. Despite recommending a personalized learning sequence, e-learning instructional strategists have failed to perform or address the necessary corrective measures to remediate immediately learning misconceptions or difficulties. As e-learning materials continue to evolve, it is necessary that an alternative, dynamic, and real time multi-performance be developed and implemented in e-learning systems. Two major contributions in the field of e-learning have been asserted by this study: it personalizes the learning sequence using reversed roulette wheel selection algorithm blended with linear ranking based on real time, dynamic multi-based performance matrix; and implements the reinforcement and mastery learning to motivate students and improve their learning output. Based on experiments, personalized learning sequence (PLS) were dynamic and heuristic and simultaneously considers the curriculum difficulty level and the curriculum continuity of successive curriculum while implementing personalized learning process. From 34%, the passing rate of the students is increased by 54% making the overall passing rate to 88%. The increase can be attributed to the reinforcement process and mastery learning where various control mechanism are implemented to guarantee learning process. Digital transcripts based on students’ perceptions and experiences positively correlate with the result of document sentiment of +.321 while theme analysis revealed a positive attitude with the extracted words in the documents such as: very happy, friends, motivate, improve, understanding, knowledge and good. Overall, the e-learning prototype were able to show an improved academic performance of the student and address different academics and social problems and allow students to study anywhere, at their own convenience whenever online learning is possible and accessible.
Thesis submitted in total fulfillment of the requirements for the degree of Doctor of Philosophy in Computer Science at Strathmore University
Reversed Wheel Roulette Selection, Personalized learning sequence, Reinforcement learning, Mastery learning, Performance matrix