Inflation Rate Prediction based on SARIMA-RF

Abstract

Accurate prediction of the inflation rate can help governments prevent social problems caused by inflation, maintain social stability, and is crucial for the economy and society. Considering the nonlinear characteristics of China's inflation rate series, this paper uses the month-on-month Consumer Price Index (CPI) data from January 2011 to March 2023 to measure the inflation rate in China. Firstly, the monthly inflation rate series from January 2011 to December 2021 is analyzed to identify significant and complex seasonal effects. Therefore, suitable SARIMA models and Random Forest (RF) models are separately constructed. In order to better capture the nonlinear features of the inflation rate series, this paper proposes a SARIMA-RF fusion model based on simple linear weighting. Using monthly data from December 2022 to March 2023, the mean absolute error and root mean square error are used as evaluation metrics for forecasting errors. The results indicate that the prediction error of the SARIMA-RF fusion model is significantly lower than that of the SARIMA model and the RF model. The proposed model can more effectively extract the nonlinear features of China's inflation rate series and accurately predict its changes. It has valuable implications for the government, society, and individuals in China.