AHSWN Home · Issue Contents · Forthcoming Papers

IOT and Big Data-Based Teaching System of Intelligent Wireless Network Classrooms in Colleges and Universities Using ECAT-PNN
Qingkai Chen, Guohui Li and Zidong Yue

Nowadays, with the development of information technology and the wave of intelligent technology, colleges and universities are exploring new teaching models to meet the needs and development of education. However, none of the prevailing technologies integrates IoT, big data, pedagogy and holistic learning experiences for intelligent classrooms. Therefore, the proposed work established an IoT and big data-based student engagement prediction framework in the intelligent wireless network classroom teaching using ECAT-PNN. Initially, data from the IoT sensor devices are transmitted to the edge device as big data through sensor node clustering, CH selection, and optimal transmission path selection. The edge device selects the video keyframes, and the environmental sensor readings are collected from the processed big data. Next, the collected environmental readings are pre-processed, while the human and face detection is carried out for the selected frames. Afterwards, the facial points are extracted, and following that, the student engagement prediction is carried out using ECAT-PNN with XAI from the extracted facial points and pre-processed data. The feedback control system provides feedback based on the predicted XAI output to enhance the teaching system. Finally, the expected production and IoT sensor device data are stored in the cloud server via dimensionality reductions using QKEDCA. Thus, the proposed methodology outperforms the other traditional approaches by attaining a better accuracy of 98.09%.

Keywords: Intelligent Teaching System, Big Data, Wireless Sensor Networks (WSN), Internet of Things (IoTs), explainable-artificial Intelligence (XAI), Expo-Convolutional Adaptive Temporal Pooling Neural Network (ECAT-PNN), Balanced Iterative Reducing Density Reshaping and Clustering using Hierarchies (BIRD-RCH), and Fractal Lion Entropy Optimization (F-LEO).