Predicting Student Performance with Adaptive Aquila Optimization-based Deep Convolution Neural Network

PREDICTING STUDENT PERFORMANCE WITH ADAPTIVE AO BASED DCNN

Abstract

Predicting student performance is the major problem for enhancing the educational procedures. A level of student’s performance may be influenced by several factors like job of parents, sexual category and average scores obtained in prior years. Student’s performance prediction is a challenging chore, which can help educational staffs and students of educational institutions to follow the progress of students in their academic activities. Student performance enhancement and progress in educational quality are the most vital part of educational organizations. Presently, it is essential for an educational organization to predict the performance of students. Existing methods utilized only previous student performances for prediction without including other significant behaviors of students. For addressing such problems, a proficient model is proposed for prediction of student performance utilizing proposed Adaptive Aquila Optimization-allied Deep Convolution Neural Network (DCNN). In this process, data transformation is initiated using the Yeo-Johnson transformation method. Subsequently, feature selection is performed using Fisher Score to identify the most relevant features. Following feature selection, data augmentation techniques are applied to enhance the dataset. Finally, student performance is predicted through the utilization of a DCNN, with a focus on fine-tuning the network parameters for optimal performance. This fine-tuning is achieved through the use of the Adaptive Aquila Optimizer (AAO), ensuring the network is poised to deliver the best possible results in predicting student outcomes. Proposed AAO-based DCNN has achieved minimal error values of Mean Square Error, Root Mean Square Error, Mean Absolute Error, Mean Absolute Percentage Error, Mean Absolute Relative Error, Mean Squared Relative Error, and Root Mean Squared Relative Error, respectively.