Exploring Types of Creative Thinking Competencies and Influencing Factors Using Machine Learning Techniques

This study aimed to classify types of creative thinking competencies among high school students and to explore the factors influencing this classification. To achieve this, the study utilized data from the 6th year of the 'Korean Educational Longitudinal Study(KELS) 2013,' focusing on first-year high school students. The study classified latent profiles based on creative thinking competencies and applied three machine learning techniques—random forests, LASSO, and ridge regression model—to predict these profiles. Additionally, it explored the factors influencing the classification of creative thinking competencies using machine learning techniques with high predictive accuracy. The results were as follows. First, students' creative thinking competency was classified into three latent profiles. Second, among the three machine learning techniques employed, random forests exhibited the highest accuracy in predicting creative thinking competency groups. Third, using random forests, the main factors affecting the classification of creative thinking competencies were identified as class attitude, academic self-concept, self-understanding, career planning, and cooperative learning. The research findings suggest a need to shift towards learner-centered instruction and instructional design that considers career maturity to enhance creative thinking competencies.