A Genetic Algorithm-Based Personalized Remedial Learning System for Learning Object-Oriented Concepts of Java

Abstract

Contribution: An online genetic algorithm-based remedial learning system is presented in order to strengthen students’ understanding of object-oriented programming (OOP) concepts by tailoring personalized learning materials according to each student’s strengths and weaknesses. Background: Prior studies on computer programming education have analyzed methods of learning OOP, and shown that teaching this topic is a challenge. A simple and personalized learning system for generating remedial learning materials would therefore be valuable, but had yet to be designed. Intended Outcomes: Students’ grasp of OOP concepts is expected to improve through study of the tailored remedial learning materials generated by the system. Application Design: Students who had previously studied OOP were recruited to test the learning system in a two-semester pre-experiment using a one-group pre-test-post-test design. The students first took a pre-test that determined their individual strengths and weaknesses in these concepts. They then read three sets of quiz-based remedial learning materials; each set was generated by the system according to the individual student’s answers in the pre-test and previous quizzes. Findings: 1) Overall, the changes between learners’ pre-and post-test scores were significant; 2) Score changes for different learners (junior, senior, low-achievement, and high-achievement learners) and for different learning styles (intensive and non-intensive) were also significant; and 3) Score changes for low-achievement learners were greater than those for high-achievement learners.