Fusion-based learning for stress recognition in smart home: An IoMT framework

Today, in order to prevent chronic stress from causing irreparable damage, it is imperative to diagnose and treat it in its early stages. Using the Internet of Things (IoT) and automated learning methods in homes, creating an intelligent environment can help identify stress-related emotions. An approach to stress detection based on metaheuristic fuzzy inference system-based learning (fMFiS-L) and emotion recognition is presented in this paper. Accordingly, our study focuses on the use of fusion learning to diagnose stress using the healthcare system and the Internet of Medical Things (IoMT) for smart homes. Music videos were shown to participants in the first stage to arouse emotional states such as anger, anxiety, and depression. Volunteers were divided into two groups, with one group practicing Reiki meditation for two weeks. The EEG signals were recorded before and after meditation, stress levels were assessed using the Likert scale, and emotions were classified using the modified fusion fuzzy inference system. In addition, a method is presented for determining the optimal parameters in the fMFiS-L structure by optimizing the innovative gunner algorithm (AIG). We conclude that Reiki meditation can significantly reduce negative emotions and stress levels in the IoMT environment of smart homes. Furthermore, the fMFiS-L architecture was evaluated for generalization of emotion recognition based on unseen EEG data. Generally, the classification of emotions produced satisfactory results, with a 92% accuracy rate.