Enhancing interior design education through data-driven internship assessment: a machine learning approach

Internships are a significant element in interior design learning, offering students practical training, industry exposure, and opportunities for skill-building. Internship programs, however, require evidence-based assessment measures to make them successful. The current study delves into the impact of internships on curricular development as perceived by the interior design students of Jordan and Bahrain. A comparative analysis examined the disparity in professional skill development, software skills, and internship performance between the students of both countries. Machine learning algorithms were employed to identify factors influencing the students’ performance to provide a more predictive internship success estimate. Decision Tree, SVM, Random Forest, Logistic Regression, KNN, and Naïve Bayes were the classification models employed to predict internship success using software skills, internship hours, and teamwork skills. Of all these outcomes, Random Forest achieved the highest (95.1%) and AUC (0.92), and this outdid all other models as the best predictor of internship success. Regression analysis also found that internship length (β = 0.45\β = 0.42β = 0.42), teamwork (β = 0.42\β = 0.38β = 0.38), and computer skills (β = 0.38\β = 0.38β = 0.38) were the predictors of internship success. The outcomes also highlight proper internships, where one can witness that longer internships, teamwork training, and computer training for interior design programs are being cited.