Deep Learning with Statistical Analysis for Stress Prediction of Remote Working IT Employees in COVID-19 Pandemic

Abstract

The COVID-19 pandemic has transformed the working environment of employees in information technology (IT) sector from traditional office environment into remote working environment. The changes in working environment, lack of physical activities, and food intake result in direct impact on physical and mental well-being. The stress among IT employees in remote working gets increased owing to the absence of proper physical workstation, and extended inactive behaviour results in high discomfort and pain. So, the advent of deep learning (DL) models assists the stress predictive procedure in understanding the pattern proficiently and delivers efficient perceptions about probable forthcoming interventions. In this view, this study develops a novel deep learning based knowledge management for stress prediction (DLKM-SP) technique among IT employees working from remote places in COVID-19 pandemic. The proposed DLKM-SP model aims to predict the stress level of the IT employees by the selection of features and optimal classification process. In addition, the DLKM-SP technique involves correlation based feature selection and principal component analysis (PCA) based feature reduction technique to choose an optimal subset of features. Moreover, attention based bidirectional long short term memory (ABiLS TM) technique was employed for the classification process for determining the proper class labels. Furthermore, arithmetic optimization algorithm is applied to improve the training process of the ABiLS TM approach. The effectiveness of the proposed model is examined using its own stress prediction dataset with numerous samples collected from IT employees. A detailed comparison study is implemented to highlight the enhanced predictive performance of the DLKM-SP approach in terms of different evaluation measures.