Khadidos Alaa O, Manoharan Hariprasath, Khadidos Adil O, N Alanazi Mohammad, Alanazi Fuhid, Selvarajan Shitharth
Department of Information Systems, Faculty of Computing and Information Technology, King Abdulaziz University, Jeddah, Saudi Arabia.
Center of Research Excellence in Artificial Intelligence and Data Science, King Abdulaziz University, Jeddah, Saudi Arabia.
Sci Rep. 2025 May 25;15(1):18176. doi: 10.1038/s41598-025-03361-5.
The process of teaching and learning during the pandemic has been evolving globally, with many institutions transforming their approaches to enhance the teaching and learning experience. Despite the presence of improved frameworks due to the varied learning capabilities of students, it remains quite challenging to analyse individual characteristic features. Consequently, this research provides clear insights into the integration of the Personalised Learning Approach (PLA) to foster effective interaction with students. However, many existing methods suggest different techniques for evaluating learners in a hybrid mode, where obtaining clear data sets can be difficult. In the teaching and learning approach, if the defined data set from experts is clear, decisions regarding the learning characteristics of students can be made in a shorter period. In the proposed method the PLA framework categorizes learners into four engagement-based clusters using a three-dimensional sensor model and machine learning classifiers. A dual-controller mechanism (master-slave) dynamically adjusts communication intervals and optimizes video transmission, reducing latency and packet loss. The methodology is validated using MATLAB-based simulations with a dataset of 1,700-5,000 learners, analyzing throughput, delay, packet loss, and cost efficiency. The test results clearly demonstrate that the PLA outperforms the conventional method, not only with the parameters mentioned above but also in terms of cost-effectiveness using master and slave controllers.
疫情期间全球的教学过程一直在演变，许多机构正在转变其教学方法以提升教学体验。尽管由于学生的学习能力各异而出现了改进的框架，但分析个体特征仍然颇具挑战。因此，本研究为个性化学习方法（PLA）的整合提供了清晰的见解，以促进与学生的有效互动。然而，许多现有方法提出了在混合模式下评估学习者的不同技术，而在这种模式下获取清晰的数据集可能很困难。在教学方法中，如果专家定义的数据集清晰，就能在更短时间内做出关于学生学习特征的决策。在所提出的方法中，PLA框架使用三维传感器模型和机器学习分类器将学习者分为四个基于参与度的类别。一种双控制器机制（主从式）动态调整通信间隔并优化视频传输，减少延迟和数据包丢失。该方法通过基于MATLAB的模拟进行验证，模拟使用了1700 - 5000名学习者的数据集，分析吞吐量、延迟、数据包丢失和成本效率。测试结果清楚地表明，PLA不仅在上述参数方面优于传统方法，而且在使用主从控制器的成本效益方面也更胜一筹。
