A Machine Learning Approach in Predicting Student’s Academic Performance Using Artificial Neural Network

Abstract

The rate at which students succeed in their academic pursuits contributes significantly to the academic achievement of their educational institutions because it is used as a measure of the institution’s performance. Many factors could be responsible for students’ academic performance and student success. Quick understanding of weak students and providing solutions to improve their performance will significantly increase their academic success rate. Educational data mining using artificial neural network plays a crucial role in determining their likely performance and helps them to initiate measures that can reposition the students’ performance in the future. This study developed a model that predicts students’ failure and success rates with the aid of a machine learning algorithm. The study sampled 720 students from three selected tertiaries institutions in Adamawa State, Nigeria. Three hundred students were selected from Modibbo Adama University, Yola, 300 students were selected from Adamawa State University, Mubi, and 120 students were selected from Adamawa State Polytechnic, Yola. The research makes use of descriptive statistics to identify the variables that likely affect students’ academic performance. The collected data were preprocessed, cleaned, and modeled using Jupyter Notebook, a Python Anaconda development platform for artificial neural to build the student’s academic performance predictive model. The neural network is modeled with 12 input variables, two layers of hidden neurons, and one output layer. The dataset is trained using the backpropagation learning algorithm. The performance of the neural network is evaluated using k-fold cross-validation. The neural network model has achieved a good accuracy of 97.36%.

Received: 14 October 2022 | Revised: 15 December 2022 | Accepted: 29 December 2022

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

Data Availability Statement

Data available on request from the corresponding author upon reasonable request.