Document

This study examined catastrophic health expenditures (CHEs) among low-income households in South Korea using machine learning methods on 2019 Korea Health Panel data. Results showed a 26.2% incidence of CHE, with AdaBoost performing best in prediction, highlighting age, chronic illness, and unemployment as key risk factors. The findings support the need for early identification and targeted support programs to reduce financial burdens and improve health protection for vulnerable groups.