aspg-28-4510-1792155043 20261016125043 American Scientific Publishing Group admin@americaspg.com American Scientific Publishing Group Journal of Artificial Intelligence and Metaheuristics JAIM 2833-5597 2026 11 1 Osama Alabedallat School of Information Communication and Technology, Bahrain Polytechnic, Isa Town, Bahrain The recent shift to online education worldwide has highlighted the urgent need to evaluate and improve students’ flexibility in the online learning environment. The paper solves that problem by creating a predictive framework that enables measuring students’ adaptability using modern techniques for advanced perception of adaptability and advanced machine learning (ML) and metaheuristic optimization methods. The research proposes an improved variant of geese’s cooperative flight behavior, called the Dynamic Greylag Goose Optimization (DGGO) algorithm, to optimize the hyperparameters of a base model, a Decision Tree (DT). A comparison of the baseline DT showed an accuracy of 93.81% and a sensitivity (True Positive Rate, TPR) of 93.14%. The DGGO optimization further improved the DT model’s performance to 97.15% accuracy and 96.86% sensitivity, with improved convergence and tuning. To be flexible to changes, the optimized model reduces classification errors and improves predictive consistency. The findings demonstrate the effectiveness of metaheuristic-based optimization for educational data mining and provide preliminary evidence that DGGO can serve as a robust, scalable, and intelligent framework to enhance predictive analytics and decision-support capabilities in the modern e-learning environment. 2026 68 82 4510 https://creativecommons.org/licenses/by/4.0/ 10.54216/JAIM.110107 https://www.americaspg.com/journal/28/article/4510 https://www.americaspg.com/storage/articles/manuscripts/01M2ZM5345CNCDMMWRXPECGCQ2.pdf