Machine Learning Algorithm to Predict Student’s Performance: A Systematic Literature Review

Abstract

One of the ultimate goals of the learning process is the success of student learning. Using data and students' achievement with machine learning to predict the success of student learning will be a crucial contribution to everyone involved in determining appropriate strategies to help students perform. The selected 11 research articles were chosen using the inclusion criteria from 2753 articles from the IEEE Access and Science Direct database that was dated within 2019-2021 and 285 articles that were research articles. This study found that the classification machine learning algorithm was most often used in predicting the success of students' learning. Four algorithms that were used most often to predict the success of students' learning are ANN, Naïve Bayes, Logistic Regression, SVM and Decision Tree. Meanwhile, the data used in these research articles predominantly classified students' success in learning into two or three categories which are pass/fail; or fail/pass/excellent.

Analysis of the Factors Influencing the Predictive Learning Performance Using Federated Learning

“…The proposed system evaluates student academic achievement using different ML-based algorithms applied to a dataset. The assessment criteria encompass model accuracy, precision, recall, and F1Score [50,51]. The results, depicted in Figure 5, showcase the aggregated testing outcomes of the models, with 10 clients executing 90 epochs.…”

One Aggregated Approach in Multidisciplinary Based Modeling to Predict Further Students’ Education

“…Machine learning algorithms in [20,21] and data mining techniques in [22,23] are considered as one overview of student performance prediction modeling for further education in both pairs of these papers. Application of machine learning in predicting performance for computer engineering students is subject of the manuscripts [24,25] and data mining techniques are applied in predicting further education for students that study medical curriculum [26].…”

Prospects and Challenges of Using Machine Learning for Academic Forecasting

“…Musso et al [10] e study successfully forecasted students' academic success one year ahead using the ANN based on cognitive and demographic traits Hudson and Cristiano [7] e results suggest that ML can generate dependable results in prediction Elhaj et al [13] e study was empirical, and it showed the ability of KNN in prediction of learning patterns of students Ahajjam et al [24] e paper provided AI-based solutions to track students' performance and was able to recommend diagnosis for the Moroccan students Pranav et al [25] e paper provided evidence on the significance of AI in management of education data and decision-making Lidia et al [26] e paper concluded that ML will be required more in the future because of the need to assist students to overcome learning difficulties and also enhance their productivity in learning Phauk and Takeo [27] e study recommended the use of the hybrid machine learning algorithm approach to solve misclassification issues and improve academic prediction accuracy Onan and Korukoglu [28] e research proposed an ensemble method to feature selection that combines the results of numerous independent feature lists generated by various features that may be used in education Onan [29] e study provided a better approach for managing students' information system via ML Hassen et al [30] e study showed that the student's success with the aid of machine learning can be monitored using their previous performance data before they engaged in the current program Ibtehal [31] e study affirmed the applicability of ML in education technology development and deployment Feders and Anders [32] ey developed a smart algorithm that assessed the teaching methods of teachers and how it affects the understanding of their lessons by students in the class taking into consideration the former knowledge of students Popenici and Kerr [33] ey examined the various implications of ML and other relevant AI-driven systems in higher education…”

“…Traditional forecasting It gives more accurate predictions with minimal loss function [10,13] Forecast errors are more likely to occur [38] e approach is more scientifically driven [26] Suffers a lot from assumptions leading to subjective conclusions at times [7] Very demanding in computation [39] Less demanding computation It is more prone to underfitting and overfitting issues [40] Less prone to underfitting and overfitting issues Focuses more on result or outcome, but silent relationships among variables.…”