GWO-MLP Based Achievement Prediction Model for English Learning With Different Learning Motivations

English achievement prediction is vital for improving the efficiency of English teaching by addressing the specific needs of students as early as possible. However, traditional English achievement prediction models are not good at the accuracy. Hence, a data-driven and AI-driven achievement prediction model is required to solve this problem. Recently, the increasing popularity of Multiple Layer Perception (MLP) and Gray Wolf Optimizer (GWO) has provided new methods for accurate prediction of students’ achievement. In terms of influencing factors, most scholars agree that learning motivation can mediate students’ learning achievement through different factors including self-efficacy and learning strategies. This paper aims to use a data-driven research schema via the questionnaire of Students’ Motivation Toward English Learning (SMTEL) and adopts an AI-driven model by Gray Wolf Optimizer-based Multiple Layer Perceptron (GWO-MLP) to predict English learning achievement with different learning motivations. Specifically, the research questions focus on (1) how the major motivational factors influence English learning achievement; (2) what the accuracy of the proposed GWO-MLP model for English achievement prediction is. The Wilcoxon signed rank test is adopted to compare GWO-MLP to other traditional and AI-driven prediction models. For the first question, the findings reveal that self-efficacy is the most influencing factor, and English learning value comes next, while performance goal exerts minimal impact on learners’ English achievement. For the second question, the results indicate that the proposed GWO-MLP model can predict English learning achievement more accurately and efficiently. Some suggestions have been recommended at the end of this research.