PREDICTIVE ANALYTICS USING DEEP NEURAL NETWORKS

Abstract

Several studies indicate that teaching methods have a significant impact on the academic performance of the students. Similarly, they also connote a substantial dependance of a student’s performance on the methods and techniques they employ in their study schedule. In this project, a model based on Deep Neural Network (DNN) has been deployed for the Predictive Analysis of a student’s performance was tested for results. Furthermore, this model also aims at recommending study methodologies to students based upon the psychometric features derived from a self-analysis questionnaire that was prepared for the same. Various factors that influence the performance of a student were identified and corresponding data was collected. This includes internal assessment scores, scores in the 10th grade and 12th grade board exams, attendance percentage for the particular semester and travel duration, availability internet connection facility, number of arrears among the others. This model was implemented on data of the student batches graduating in years in the duration of 2016-22. Test data evaluation presents an accuracy of 96.3% in the prediction of the student performance.