What determines Student Employability? Educational Data Mining through Machine and Deep Learning Approach

Title:

What determines Student Employability? Educational Data Mining through Machine and Deep Learning Approach [Descargar]

Autores/as:

Tariq, Rasikh and Vargas, Diego Gutiérrez and Ali, Farhan and Gonzalez-Mendoza, Miguel and Torres-Castillo, Cristina Sofia

Términos Índice:

Data mining;Predictive models;Employment;Education;Machine learning;Biological system modeling;Deep learning;Data models;Decision trees;Artificial intelligence;Educational Data Mining;Student Employability;Machine and Deep Learning;Explainable AI;Educatio

Resumen:

Employability is vital for graduates to succeed in competitive job markets and reflects higher education institutions' effectiveness. It is essential to investigate which specific traits contribute to a higher success rate of employability, as understanding these factors can help optimize targeted interventions and improve employment outcomes. The objective of this research is to identify and analyze the key traits that influence student employability using educational data mining techniques integrated with machine learning and deep learning models while providing an explainable framework to inform targeted interventions and enhance job market readiness among graduates. Addressing gaps in existing research, this study integrates a wide range of variables and employs advanced Artificial Intelligence (AI) techniques, specifically Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU), to develop a predictive framework for understanding employability over time. Using data from mock job interviews, the study applies Shapley Additive exPlanations (SHAP) values to assess the impact of traits like Self-Confidence and Ability to Present Ideas. Hyperparameter tuning through Grid Search and k-fold cross-validation is employed to optimize model performance. The LSTM model, configured with three layers, achieved an accuracy of 91.46%, and demonstrated the highest performance among the evaluated models. Its robustness was further supported by a 90.48% accuracy obtained through 3-fold cross-validation. The current findings highlight the importance of soft skills, such as Self-Confidence, Ability to Present Ideas, and General Appearance, identified by SHAP analysis as critical predictors of employability, emphasizing the need for educational institutions to actively integrate soft skills development into their curricula to ensure students are both academically prepared and professionally equipped.

DOI:

10.1109/RITA.2025.3612280

Cómo citar:
Tariq, Rasikh and Vargas, Diego Gutiérrez and Ali, Farhan and Gonzalez-Mendoza, Miguel and Torres-Castillo, Cristina Sofia, "What determines Student Employability? Educational Data Mining through Machine and Deep Learning Approach" in IEEE Revista Iberoamericana de Tecnologias del Aprendizaje, pp. 1-1, . 2025. doi: 10.1109/RITA.2025.3612280