Relative Comparison of Healthcare Supply

Utilizing Spatial Machine Learning for the Relative Comparison of Healthcare Supply across Local Governments

지방정부 의료공급의 상대적 비교를 위한 공간정보 기반 머신러닝의 활용

Read Paper here

Authors: Dohyo Jeong

Abstract

Differences in healthcare supply across regions have been widely discussed in relation to how local governments allocate healthcare resources. Previous studies have focused mainly on healthcare resource distribution and regional disparities, while relatively little attention has been given to assessing healthcare supply in relation to local characteristics. This study used machine learning incorporating spatial information to predict the expected healthcare supply for local governments. The predicted expected healthcare supply was compared with the observed healthcare supply to examine relative regional deviations. Explainable machine learning was also applied to identify the local characteristics that contributed most to healthcare supply predictions and to examine how predictions responded to different input values. The results showed that even regions with similar observed healthcare supply could differ in their relative position compared with expected healthcare supply, depending on local characteristics. Prediction responses to input characteristics also varied across regions. This study contributes by using explainable machine learning with spatial information to jointly examine relative deviations in healthcare supply and the key characteristics used in prediction.

Key Words: Machine Learning, Explainable Artificial Intelligence, Healthcare Supply, Local Government, Spatial Information