Predicting Post-Internship Employability Using Ensemble Machine Learning Approach

Aiza Azlin, Kahlik and Abdulrazak Yahya, Saleh (2024) Predicting Post-Internship Employability Using Ensemble Machine Learning Approach. Journal of Cognitve Sciences and Human Development, 10 (2). pp. 87-101. ISSN 2550-1623

PDF Predicting Post-Internship Employability Using Ensemble Machine.pdf Download (258kB)

Abstract

Graduate employability is crucial for both students and higher education institutions. While academic performance has traditionally been a key predictor of employability, its predictive power is limited, necessitating the exploration of additional factors influencing post-internship job placement. This study investigates the impact of internship-related variables on graduate employability, such as duration, training performance, and prior work experience. Employing a machine learning approach on a dataset comprising student records from Universiti Malaysia Sarawak spanning from 2019 to 2021, we compared the performance of various algorithms, including ensemble methods. Feature selection and repeated K-fold cross-validation optimised model performance. Results indicate that stacking outperforms traditional models, achieving an accuracy of 91%. Particularly, internship duration and training performance emerged as significant predictors of employability. These findings underscore the importance of robust internship programs in enhancing graduate outcomes. Future research could explore the competencies developed during internships and their correlation with job success.

Actions (For repository members only: login required)

View Item