Bitcoin Price Forecasting Based on ARIMA Model and Multifactorial Linear Regression

Contemporarily, Bitcoin has enjoyed great popularity worldwide based on the development of the blockchain technology. Dozens of researches have been done on forecasting the price of bitcoin, however, the most accurate model remains inconclusive. This paper investigates the performances of Bitcoin price prediction based on ARIMA model and ordinary least square multifactorial linear model in terms of the dataset from May 2019--April 2022 in both short-term and long-term. According to the analysis, linear regression model outperforms ARIMA model in short term test which is significantly accurate. Nevertheless, in terms of long-term prediction, although the linear regression model still performs better than ARIMA model. None of them shows great accuracy on account of uncertain changes that some assumptions are no longer applicable. Therefore, these results shed light on guiding further exploration focusing on cryptocurrency pricing.