Evaluating Student Performance Based on Deep Learning Predictions

Artificial intelligence has been widely used and attracted many researchers in the field of education due to its results and performance and has increased the demand for student performance evaluation by providing data-driven insights that can improve educational decision-making. Traditional assessment methods have limited ability to manage the complex and interconnected factors that influence academic success, and because of this, machine learning can improve institutions ability to identify students at risk of failure and make correct interventions that improve student outcomes and evolve and integrate into educational practices. As a result, it enhances the learning experience, supports student success, and improves overall educational processes. In this paper, the student performance dataset serves as a synthetic representation of various factors that influence academic success. The dataset includes several features such as study time, sleep hours, socioeconomic background, and class attendance, all of which have a direct or indirect impact on student results. The main goal of this paper is to predict whether a student will pass or fail based on these features and solve a complex real-world problem using machine learning. Trained the proposed model on this dataset, which can serve as a tool for educators and institutions to understand the factors that contribute to student success and to predict academic outcomes. Finally, the model with the best accuracy of 97.62% and average accuracy of 95% has proven to be efficient in predicting student success in the future, and by using our method, the institutions can develop targeted interventions to help struggling students, ultimately improving educational outcomes.