Will the Student Get an A Grade? Machine Learning-based Student Performance Prediction in Smart Campus: Fid-Bau Portal

Will the Student Get an A Grade? Machine Learning-based Student Performance Prediction in Smart Campus

Alnoman, Ali

As an important component in the structure of smart campus, student performance prediction can help to observe the academic progress of students and make timely decisions towards improving the overall learning process. In order to achieve accurate performance prediction, it is required to extract useful data such as the past performance of students from the student information system and employ the extracted data to train an appropriate machine learning model. In this paper, different machine learning models are utilized to predict whether or not a student will get an A grade (i.e, 90 or higher). The work utilizes a real dataset that is extracted from three different courses that require computer and programming skills. The dataset includes three features about the student past performance, namely, high school grade, course midterm grade, and absence percentage. The dataset is then used to train different machine learning models, specifically, linear discriminant, logistic regression, Naive Bayes, support vector machine, decision tree, K-nearest neighbors, and bagged trees. In order to highlight the effectiveness of these classifiers, different metrics were used to evaluate the classification performance such as accuracy, precision, recall, and F1-score. Besides, these models are tested considering one, two, and three features from the dataset to evaluate the significance of each feature in the classification process. After comparing the performance of these machine learning models, it is shown that predicting student performance is indeed applicable with an accuracy that reached 99%. It is also shown that the bagged trees and K-nearest neighbors succeeded to achieve the highest classification accuracy compared to the other models.