A Proposal for a Deep Learning Model to Enhance Student Guidance and Reduce Dropout

Abstract

Despite attempts to improve the university’s training offer, dropout rates are constantly increasing. This paper is an attempt to optimize university’s training offer that would foster students’ guidance thanks to a neural network model. The study consists in collecting data from university’s database about students with regard to their gender, province, institution, course of study, grades, validated modules/semesters, success and graduation rates, etc. These data were collected and analyzed to make predictions on any kind of student based on the proposed deep learning model. These predictions allow proposing the suitable course to follow by each future student. The beginning of our work concerns the collection of the dataset which we used in deep learning model as training data. Afterward, we proposed a deep learning model that optimizes our university’s training offer. The goal of the present work is to increase the student success rates by providing them with the most appropriate course in the university using the prediction tools. Thus, it is hoped that this study will contribute to decrease the rates of students’ dropout.