Design of Student Behavior Analysis and Management System Based on Deep Learning

The purpose of the research is to utilize Deep Learning (DL) algorithms to analyze student behavior and learning management systems, to address the growing need for effective monitoring and intervention strategies in educational environments. This study proposed a novel parallel-computing fruit fly-optimized adjustable recurrent neural network (PFFO-ARNN) to analyze student online temporal behaviors for the early prediction of course performance. For this research, student behavior data was collected from the learning management system used in the college. The data is preprocessed using tokenization and the features are extracted using word2vec to convert words into numerical vectors from preprocessed data. We evaluate the prediction performance of PFFO-ARNN using various parameters and compare it to other conventional techniques. The findings show that early identification of students at risk with a greater level of prediction Accuracy, Precision, Recall, and F1-score was achieved by employing the PFFO-ARNN. Furthermore, compared to other conventional techniques, the PFFO-ARNN approach demonstrated greater generalization and better prediction performance.