Knowledge Refinement Tools to Support Inductive Learning by Inquiry Examples

Abstract

This paper presents a machine learning approach to develop computer-assisted learning supports for inductive learning tasks. Several machine-learning techniques are developed to evaluate student’s learning results, promote learning and achieve better results by giving better hints during the learning process. The learning supports can be exploited to provide individualized guidance in the context of learning classification knowledge by inquiry examples. Integrated with the learning supports, a knowledge refinement process is proposed, and a web-based system, named ALBIX (Active Learning By Inquiry Examples), was implemented so that students can actively construct, verify and refine their classification knowledge in an interactive manner. The knowledge refinement process is supported by four machine-learning modules, which are knowledge retraction module, knowledge evaluation module, knowledge diagnosis and knowledge remediation module, respectively. Finally, the learning supports presented in this paper are shown to be effective to their design purposes through a set of simulation tests. A small-scaled prototype testing also showed that teachers and students might be interested in such a kind of learning strategy.