A WEB-BASED STUDENTS’ ATTENDANCE PREDICTION SYSTEM USING MACHINE LEARNING

Abstract

This study details the creation of a web-based system designed to predict student attendance using machine learning. The main goal is to provide educational institutions with accurate attendance predictions to help address potential issues early. The system used past two years attendance from the school of electrical systems engineering at the Federal University of Technology, Akure, Nigeria. The dataset comprises of records of students’ lectures attendance of four departments for two academic sessions. These records show students response to classes for the five working days of the week. The preprocessed data was used to train and evaluate three machine learning models; Linear Regression, ARIMA, and XGBoost. The research was implemented on Python independent development environment for data processing and model development. Flask was employed as the web framework, and HTML/CSS for the front-end design. Visual Studio Code (VS Code) was the chosen environment for writing and debugging the code. After testing the models, Linear Regression showed the best results due to its ability to model the data with a linear relationship. The web platform allows users to input data and receive attendance predictions, providing a useful tool for educators and administrators. Overall, this study highlights how machine learning can be applied to education management, offering a scalable solution for predicting attendance, which can support better planning and student engagement.