Employability Prediction Based on Student Performance Using CNN Model

Unemployment among graduates is a persistent issue, largely because traditional academic evaluations don't always reflect the skills employers seek. This study explores a new approach to predict employability based on students' exam performance using Convolutional Neural Networks (CNNs). We utilized the “Student Performance” dataset, which includes a variety of academic, demographic, and socioeconomic factors. By comparing CNNs with other machine learning models such as SVM, Decision Trees, and XGBoost, we observed that SVM obtained an accuracy of 94.33%, whereas Decision Trees and XGBoost obtained 96.00% and 97.00%, respectively. CNN obtained the highest accuracy of 99.50%. CNN is sometimes misclassified but its strength lies in unearthing complex patterns in student data. This research underscores how advanced machine learning models, especially CNNs, can enhance employability predictions, enabling educational institutions to refine their curricula and better prepare students for the evolving job market.