Predicting Mental Health Issues Among Employees Using a Hybrid Logistic Regression and Random Forest-Based Monitoring System

Mental health is now a world concern that affects significantly the personal and working life of most individuals. The World Health Organization (WHO) concluded that over 20 % of the world population suffers from mental illnesses, while one in every five workers feels stress, anxiety, or depression that leads to low workplace productivity. Mental illness remains undiagnosed due to stigma, ignorance and lack of organizational support systems. In response to this, machine learning (ML) has exhibited tremendous potential to revolutionize the early detection, diagnosis, and treatment of mental illnesses. Using large-scale psychological and behavioral data, machine learning algorithms are capable of revealing underlying patterns and informing timely interventions. This article delves into the use of machine learning to deter-mine employee mental health through data-driven analytics. We evaluated data from the Open Sourcing Mental Illness (OSMI) Mental Health in Tech Survey and recommend a hybrid machine learning model combining Logistic Regression (LR) and Random Forest (RF) classifiers. The model aims to improve prediction accuracy and alleviate overfitting, allowing companies to classify high-risk employees with higher accuracy. The experimental re-sults show that our hybrid approach achieves an accuracy of 97 % compared to the isolationist ML models. The evidence confirms the increasing phenomenon of using artificial intelligence to treat mental illnesses and provides a new paradigm to help improve corporate wellness. Finally, the research indicates a synergy between machine learning and mindfulness and recommends its utilization in contemporary work environments.