COMBINING MACHINE LEARNING AND STATISTICAL MODELS IN TIME SERIES FORECASTING: CASE OF INFLATION IN VIETNAM PERIOD 2000 - 2021

Abstract

Time series forecasting is a very important problem in production, business and policy making. In Vietnam, many studies have used statistical models and deep learning models independently to forecast time series such as: amount of foreign investment, stock index, consumer price index, etc. However, the combination of the above models in forecasting economic variables is still quite rare in Vietnam. The article aims to determine the optimal combination between deep learning models and traditional machine learning models in forecasting the inflation index of Vietnam in the period 2000 - 2021.