A Personality-Driven Recommender System for Cross-Domain Learning Based on Holland Code Assessments

Abstract

Over the past few decades, AI has been widely used in the field of education. However, very little attention has been paid to the use of AI for enhancing the quality of cross-domain learning. College/university students are often interested in different domains of knowledge but may be unaware of how to choose relevant cross-domain courses. Therefore, this paper presents a personality-driven recommender system that suggests cross-domain courses and related jobs by computing personality similarities and probable course grades. In this study, 710 students from 12 departments in a Taiwanese university conducted Holland code assessments. Based on the assessments, a comprehensive empirical study, including objective and subjective evaluations, was performed. The results reveal that (1) the recommender system shows very promising performances in predicting course grades (objective evaluations), (2) most of the student testers had encountered difficulties in selecting cross-domain courses and needed the further support of a recommender system, and (3) most of the student testers positively rated the proposed system (subjective evaluations). In summary, Holland code assessments are useful for connecting personalities, interests and learning styles, and the proposed system provides helpful information that supports good decision-making when choosing cross-domain courses.