Predicting Student Performance with Secure Data Handling and Deep Learning-Based Classification Models

Predicting student academic performance is a critical task for educational institutions, as it enables early identification of at-risk students and supports informed academic decision-making. This study proposes a binary classification framework for predicting student performance using deep learning techniques, specifically Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM) models.
The proposed approach formulates the prediction task as a classification problem, where students are assigned to predefined performance categories rather than predicting exact numerical scores. Prior to model training, data preprocessing and feature standardization are applied to enhance learning efficiency. To ensure data confidentiality and integrity, the Advanced Encryption Standard (AES) is employed to encrypt the student dataset before storage.
 The performance of the proposed models is evaluated using standard classification metrics, including accuracy, precision, sensitivity, and F1-score. Experimental results demonstrate that the RNN model achieves superior performance with an accuracy of 96%, outperforming the LSTM model, which attains an accuracy of 75%. These findings highlight the effectiveness of deep learning models for student performance classification and emphasize the importance of secure data handling in educational prediction systems.