Development and Evaluation of Adaptive LearningSupport System Based on Ontology of MultipleProgramming Languages

TL;DR

This paper presents ADVENTURE, an adaptive programming learning system using an ontology across multiple languages and the Elo rating for personalization, showing improved learner performance over random modes.

Contribution

Introduces a novel ontology-based adaptive learning system for programming that personalizes exercises and hints across multiple languages, evaluated through experimental comparison.

Findings

Significant improvement in correct submissions with adaptive mode

Adaptive mode leads to more pass concepts

System effectively supports programming practice

Abstract

This paper introduces an ontology-based approach within an adaptive learning support system for computer programming. This system (named ADVENTURE) is designed to deliver personalized programming exercises that are tailored to individual learners' skill levels. ADVENTURE utilizes an ontology, named CONTINUOUS, which encompasses common concepts across multiple programming languages. The system leverages this ontology not only to visualize programming concepts but also to provide hints during practice programming exercises and recommend subsequent programming concepts. The adaptive mechanism is driven by the Elo Rating System, applied in an educational context to dynamically estimate the most appropriate exercise difficulty for each learner. An experimental study compared two instructional modes, adaptive and random, based on six features derived from 1,186 code submissions across all the experimental groups. The results indicate significant differences in four of six analyzed features between these two modes. Notably, the adaptive mode demonstrates a significant difference over the random mode in two features, the submission of correct answers and the number of pass concepts. Therefore, these results underscore that this adaptive learning support system may support learners in practicing programming exercises.