Evaluating and Forecasting Student Academic Performance through Educational Data Mining using Deep Learning

Abstract

Frequent failures during various stages of education are a common occurrence. Numerous factors contribute to the increasing drop-out rates among students, with poor academic performance being a significant reason for school discontinuation. Many students struggle to adapt to the academic environment of their institutions, which negatively impacts their performance. Additionally, involvement in extracurricular activities and student politics often diverts focus, leading to unsatisfactory outcomes. These predictable and unforeseen factors collectively influence academic growth and development. Consequently, it is essential to analyze undergraduate performance to uncover the underlying causes of students' varying achievements. The main objective of our study is to detect the diverse factors that impact academic success at the undergraduate level. The ultimate goal of this study is to empower students by helping them understand these influencing factors, enabling them to take proactive steps to enhance their academic results. By identifying and assessing these key elements, students, educators, and institutional stakeholders can work collaboratively to create a more conducive learning environment. This paper emphasizes the critical role of utilizing student data in improving education planning. It outlines effective techniques to extract valuable insights from extensive academic databases containing student information. A deep learning-based Recurrent Neural Network classifier model is suggested to help make early predictions about how well students will do in school. The proposed approach is associated against many traditional machine learning classification models and RNN classifiers to evaluate its efficacy.