Genetic Algorithm-Based Human Mental Stress Detection and Alerting in Internet of Things. (31st August 2022)

• Record Type: Journal Article
• Authors: Hamatta, Hatem S. A.; Banerjee, Kakoli; Anandaram, Harishchander; Shabbir Alam, Mohammad; Deva Durai, C. Anand; Parvathi Devi, B.; Palivela, Hemant; Rajagopal, R.; Yeshitla, Alazar
• Other Names: Kumar Vijay Academic Editor.
• Is Part Of: Computational intelligence and neuroscience. Volume 2022(2022)
• Publication Date: 2022-08-31
• Subjects: Neurosciences -- Data processing -- Periodicals; Computational intelligence -- Periodicals; Computational neuroscience -- Periodicals; 612.80285
• DOI: 10.1155/2022/4086213
• Languages: English
• ISSNs: 1687-5265

Abstract

Abstract : Healthcare is one of the emerging application fields in the Internet of Things (IoT). Stress is a heightened psycho-physiological condition of the human that occurs in response to major objects or events. Stress factors are environmental elements that lead to stress. A person's emotional well-being can be negatively impacted by long-term exposure to several stresses affecting at the same time, which can cause chronic health issues. To avoid strain problems, it is vital to recognize them in their early stages, which can only be done through regular stress monitoring. Wearable gadgets offer constant and real information collecting, which aids in experiencing an increase. An investigation of stress discovery using detecting devices and deep learning-based is implemented in this work. This proposed work investigates stress detection techniques that are utilized with detecting hardware, for example, electroencephalography (EEG), photoplethysmography (PPG), and the Galvanic skin reaction (GSR) as well as in various conditions including traveling and learning. A genetic algorithm is utilized to separate the features, and the ECNN-LSTM is utilized to classify the given information by utilizing the DEAP dataset. Before that, preprocessing strategies are proposed for eliminating artifacts in the signal. Then, the stress that is beyond the threshold value is reached the emergency/alert state; in that case, an expert who predicts the mental stress sends the report to the patient/doctor through the Internet. Finally, the performance is evaluated and compared with the traditional approaches in terms of accuracy, f1-score, precision, and recall.