Hybrid Arimax-NN Model for Forecasting Inflation

Abstract

Inflation became an important component in the economy as an indicator of the increase in prices of goods and services. In addition to general inflation, there are also seven groups of inflation categorized based on expenditure. Inflation particularly in Indonesia is influenced by internal and external factors. These factors may effect inflation not only at a single point of time, but also at certain periods. Money supply is one factor strongly considered to influence inflation. Consequently, it is important to forecast inflation by involving money supply as input series. The effect of money supply on inflation was analyzed in this study. This research focused on hybrid method which is the combination between Autoregressive Integrated Moving Average with Exogenous Factor (ARIMAX) and Neural Network (NN). The results of hybrid method were compared to individual forecasting method, i.e. ARIMA and ARIMAX. The result indicated that hybrid ARIMAX-NN provided precise inflation prediction compared to ARIMA or ARIMAX method. Hybrid model can be an effective and efficient way to improve forecasting.