Forecasting Cryptocurrencies using the Classical Time Series Approach

Abstract

As technology leads toward a new era of tools, there are also some sudden changes in business and marketing. Cryptocurrency is a new and emerging investment and exchanging tool for business and marketing. This work aims to develop a time series model that efficiently forecasts cryptocurrency values. To achieve this end, we use the classical time series model Autoregressive integrated moving average (ARIMA), also known as the Box-Jenkins methodology. This work demonstrates that by using ARIMA models, the future behavior of the series can be efficiently guessed. The work suggests some ARIMA models by utilizing the Box-Jenkins methodology that can efficiently guess the future behavior of the cryptocurrency, and these models are selected based on forecast accuracy. Namely root mean square error (RMSE), mean absolute percentage error (MAPE), and the Akaike information criteria (AIC). The results show that different models are selected to model and forecast the four cryptocurrencies. These results will provide an initial guess to the investors and consumers to know the behavior of the cryptocurrencies in the upcoming days.