머신러닝 접근의 재정관리: 세입추세 예측모형 연구 = Revenue Forecasting with Machine Learning Approach

다국어 초록 (Multilingual Abstract)

Fiscal sustainability received much attention due to recent pandemic and countercyclical fiscal policy. This study attempts to apply the logic of machine learning algorithms to financial management for policy implications. Practically, the growing evidence has documented the increasing use of cases with machine learning algorithms. However, there are limited studies with in scholarly works from the field of public administration. Using 50 years of revenue data from Seoul metropolitan area and 20 years of 69 local governments, our findings reveal that exponential smoothing works better for Seoul metropolitan area, while KNN is superior to revenue forecasting in other local governments.