Comparative Analysis of ARIMA and Random Forest Regression Models in Predicting Money Supply Trends

Abstract

This paper collects monthly data of money supply over an extended period, including eight indicators: money and quasi-money, its year-on-year growth, currency and its year-on-year growth, cash in circulation and its year-on-year growth, current deposit and its year-on-year growth. Two distinct modeling approaches, the ARIMA model and the random forest regression model, are employed to analyze these data. In the ARIMA model, only the money and quasi-money data are utilized, whereas in the random forest regression model, money and quasi-money are treated as dependent variables with the other indicators serving as independent variables. Despite the variations in data usage between the two models, their predictive performances are comparable.