Forecasting the Volume of Electronic Money Transactions Using ARIMAX-GARCH Model and Support Vector Regression

Increasing economic activity is part of the Sustainable Development Goals (SDGs) in encouraging economic growth. This encourages the transformation of the digital economy that gave birth to a payment system with electronic money. The use of electronic money has many benefits such as accelerating the circulation of money and providing convenience in every economic and trading activity. In Indonesia, the volume of electronic money transactions has increased significantly in recent years. For this reason, it is very important to forecast the volume of electronic money transactions in Indonesia as a material for evaluating policies to increase economic growth through the role of the digital economy. This is the aim of this research. In this study, the forecasting of the volume of electronic money transactions is carried out by considering several exogenous variables such as electronic money infrastructure and the conditions of the COVID-19 pandemic. The study was conducted by comparing the accuracy of the model obtained from the ARIMAX-GARCH model and Support Vector Regression (SVR) to predict the volume of electronic money transactions in Indonesia. The results showed that the two exogenous variables had a significant effect on the volume of electronic money transactions. In addition, the SVR model shows better accuracy than the ARIMAX-GARCH model which is indicated by the Mean Absolute Percentage Error (MAPE) value of the testing data of 7.01% and 3.35%, respectively. Thus, this research is expected to be used as a basis for formulating policies related to the development of the digital economy, especially the use of electronic money to support economic growth in Indonesia.