Predicting employees' coping behaviors against job stress using a hybrid modeling approach based on structural equation analysis and machine learning

Background and Objectives: This research aimed to predict employees' coping behaviors when faced with job stress, using a combination of structural equation analysis and machine learning methods. This study seeks to provide a scientific framework for better understanding stress coping mechanisms in workplaces. Methodology: The statistical population consisted of 240 employees of a service organization who were selected by simple random sampling method. Data were collected through standardized questionnaires and their reliability was confirmed with a Cronbach's alpha coefficient above 0.85. The data analysis process was carried out in two main stages: in the first stage, structural equation analysis (SEM) was used to examine the relationships between variables, and in the second stage, modeling was carried out with various machine learning algorithms. Findings: The results of structural equation analysis showed that the three variables of job stress, organizational support, and self-efficacy have a significant effect on coping behaviors. In the machine learning section, the multilayer neural network (MLP) algorithm was recognized as the best predictive tool with a significant accuracy of 92.5% and a coefficient of determination of 0.68, which indicates its high predictive power. Conclusion: The hybrid model presented in this study is considered an efficient, accurate and scientific tool for predicting coping behaviors and designing stress management programs in various organizations. This model can help managers and decision-makers to better understand employees' reactions to stress and implement more effective strategies to help improve mental health and increase efficiency in the workplace.