Understanding psychological distress : a predictive model using machine learning's classification trees

Abstract

When the global COVID-19 pandemic saw its outburst, there was a debate whether it was best to impose restrictions, namely quarantine. However, one of the counter-arguments fell around psychology: limited mobility and for many, temporary unemployment, could present a major challenge for personal well-being. Given the complex nature of individual mental health, the purpose of this paper is to understand some of the features that lie behind psychological distress, and how the presence or absence of some indicators may be key to comprehend the latter better. Using NHIS information, a predictive model has been generated by the Random Forest algorithm in the Python environment. The results show that physical restrictions, financial limitations, and sleep quality, amongst others, have a major relevance in understanding psychological distress.