A Comprehensive Computational Framework for Detecting and Analyzing Human Stress in Workplace Through Rough Set Theory and ICF

In today’s fast-paced world, an important problem is the acute level of stress spread out in workplaces. This phenomenon can have heavy consequences on the individuals’ health. In this paper, we outline a comprehensive framework using machine learning, situational awareness, and rough set theory. We first analyze how to detect the stress level of workers using machine learning, avoiding intrusive equipment. Secondly, we carry on an analysis on how to prevent the spread of acute stress levels among a population of workers. For this purpose, we employ the Situational Awareness theory and, on its principles, design a decision dashboard that can be helpful for this goal. Next, we put emphasis on how to find the root causes of the stress by defining a methodology that combines ICF (a de-facto standard for describing levels of human functionality) with Rough Set theory. Our results show how the Random Forest is the model that performs better using the EDA and ECG signals from the WESAD dataset and how to engineer the making of the predictions in a real-time environment. Moreover, we show how practical use of the ICF qualifiers and Rough Set theory can be useful in the deduction of a root cause.