Effective Stress Detection and Classification System Using African Buffalo Optimization and Recalling-Enhanced Recurrent Neural Network for Nano-Electronic Typed Data

Abstract

A body’s altered emotional reactions to a variety of conditions, including despair, anxiety, rage, grief, guilt, low self-worth, etc., can lead to stress. Stress hurts a person’s performance and is the underlying cause of many mental health issues, including dementia and depression. Numerous prevailing approaches to stress detection are exploited with deep learning, but it needs to categorize the stress precisely, and it takes high computation time. To engulf these complications, an African buffalo optimization and the Recalling-Enhanced Recurrent Neural Network (RE-RNN) are newly proposed for accurately detecting stress. At first, the stress dataset is collected from the Kaggle website, which actually hold the records for the data generated using the nanoelectronic and optoelectronic devices. Afterward, the preprocessing method eliminates noise and improves input data by utilizing adaptive filter method. Next, the preprocessing output is fed to the Feature extraction section. The features are extracted based on discrete wavelet Transform (DWT). After that, the extracted data are updated to the classification process using a Recalling-Enhanced Recurrent Neural Network (RE-RNN) to accurately detect stress. Hence, the African Buffalo Optimization (ABO) is proposed to adjust RE-RNN, which precisely classifies stress detection. The performance of the proposed RE-RNN approach attains 99.89%, 98 98.76 and 98.07% high accuracy, and 0.1%, 0.2%, and 0.2% lower computation Time.