English Language Teaching; Vol. 16, No. 4; 2023 ISSN 1916-4742 E-ISSN 1916-4750 Published by Canadian Center of Science and Education 24 Prediction of Students’ Performance in English Using Machine Learning Algorithms Nipa Jun-on 1 , Pimpaka Intaros 1 & Sarawut Suwannaut 1 1 Department of Mathematics, Faculty of Science, Lampang Rajabhat University, Lampang 52100, Thailand Correspondence: Sarawut Suwannaut, Department of Mathematics, Faculty of Science, Lampang Rajabhat University, Lampang 52100, Thailand. E-mail: sarawut-suwan@hotmail.co.th Received: January 1, 2023 Accepted: March 18, 2023 Online Published: March 20, 2023 doi: 10.5539/elt.v16n4p24 URL: https://doi.org/10.5539/elt.v16n4p24 Abstract In this work, a new machine learning-based model is proposed to predict undergraduate students' reading scores using their listening scores as the primary data. The performance of several machine learning techniques, including neural networks, gaussian process regression, and random forests, was calculated and compared in order to predict the reading test results of the students. The dataset included the listening and reading test results of 1145 students who took the English proficiency exam at Lampang Rajabhat University's language center in Lampang, Thailand. According to the results, the suggested model has a classification accuracy range of 64–75%. Only three different types of parameters—listening scores, departmental data, and faculty data—were used to make the predictions. Keywords: machine learning algorithms, educational data mining, prediction, English 1. Introduction Education data mining (EDM) is the use of data mining (DM) techniques on data acquired from various educational systems with the goal of improving education (Baker & Yacef, 2009). Recent EDM research mainly concentrates on assessing students' learning and behavior, analyzing educational tactics and interventions, predicting students' performance and dropout, and giving students tailored suggestions (Romero & Ventura, 2020). One of the top considerations for educators in assessing educational achievement at all educational levels is students' academic performance (Tan et al., 2019). Educators can develop and execute early interventions and support to improve students' performance when their performance is predicted, especially for students who are at risk of failure (Wakelam et al., 2020). Even while DM is becoming more and more common in educational settings, EDM studies are still quite few and are not well covered, especially in developing nations like Thailand. The limited number of prediction-related EDM research conducted in the Thai education area (Pattanaphanchai et al., 2019; Iam-On & Boongoen, 2017). As a result, relatively little is known about the factors that predict academic achievement for students in this area. The most common activity in educational data mining to create a prediction model of student performance is classification. To forecast students' performance, a variety of classification approach algorithms were used, including Decision Tree, Neural Network, Naive Bayes, and Support Vector Machine (SVM) (Shahiri et al., 2015). A decision tree is used to find useful information in a small or large data collection. Decision tree algorithm is a popular prediction technique because it is easy to understand. In a given data collection, Naïve Bayes computes a set of probabilities in a given data set (Patil & Sherekar, 2013). A supervised machine learning approach called Support Vector Machine locates the hyperplane in an N-dimensional space that optimally separates the data into two categories (Patil & Sherekar, 2013). EDM provides new knowledge to educators by identifying hidden patterns in educational data. Many components of the educational system may be evaluated and changed using this method to guarantee the quality of education. 2. Literature Review Numerous studies have been conducted on forecasting students’ performance. The methodology for predicting a student's grade point average at graduation was presented by Tekin (2014). The prediction model was created