Abstract

The technology sector faces a critical operational gap in proactively managing workforce mental well-being due to the pervasive stigma surrounding treatment-seeking. This study addresses the challenge of predicting whether tech professionals will voluntarily seek mental health support by leveraging machine learning techniques. Analyzing over 1,200 responses from the OSMI Mental Health in Tech Survey, this research models the influence of demographic factors, workplace benefits, and anonymity on care-seeking behavior. A comprehensive preprocessing pipeline was implemented, utilizing SMOTE to address class imbalance and benchmarking twelve classification algorithms. Results demonstrate that the Random Forest classifier achieves the highest performance with an accuracy of 83.07% and an <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$F 1$</tex>-score of 0.83, effectively capturing nonlinear interactions within the data. Notably, the Random Forest ensemble outperformed the single Decision Tree baseline by approximately 7.6%, validating the necessity of ensemble methods in reducing variance. These findings highlight the potential of predictive analytics in enabling organizations to design targeted, privacy-preserving mental health interventions.

Concepts :

Access to Document

Citations by Year