Comparison of Forcasting Ability between Backpropagation Network and ARIMA in the Prediction of Bitcoin Price

Abstract

Bitcoin is a peer-to-peer (P2P) electronic currency that allows online payments around the world without the management of a third party. Many studies have been conducted on the performance prediction of the stock market; in particular, the Autoregressive Integrated Moving Average model (ARIMA) is one of the linear regressive models widely used in the time series. Nevertheless, as artificial intelligence has become a heated research topic in modern days, A number of studies have also shown that the Back-propagation Neural Network (BPNN) is very effective in prediction. Hence, this paper compares the ARIMA model with the BPNN model in the prediction of Bitcoin price.