ENOP: A Robust Methodology to Predict Working Employee Stress Levels Using Enhanced Neural Optimization Principle
Abstract
Employee stress is a significant concern in modern workplaces, especially in high-demand industries like IT, where prolonged stress can lead to mental health issues and reduced productivity. This paper introduces ENOP (Enhanced Neural Optimization Principle), a robust methodology to predict employee stress levels using a hybrid approach combining AlexNet for feature extraction and LightGBM for classification. The model leverages physiological data such as Galvanic Skin Response (GSR) and Electrocardiogram (ECG) along with survey-based demographic and psychological factors. Key features include health concerns, pressure, financial and family issues, working hours, and regular interaction. The AlexNet-LightGBM model demonstrated high performance across multiple metrics, achieving an accuracy of 92%, precision of 89%, recall of 91%, and an F1-score of 90%. This hybrid approach outperforms traditional machine learning methods in both processing efficiency (92%) and training time (280 seconds), making it suitable for real-time stress monitoring and management. The model's ability to provide actionable insights into stress-inducing factors allows for targeted interventions, helping improve workplace productivity and employee well-being.