Forecasting Cryptocurrencies Volatility Using Statistical and Machine Learning Methods: A Comparative Study

Abstract

This paper presents a comprehensive study of statistical and machine learning methods for predicting daily and weekly volatility of the following four cryptocurrencies: Bitcoin, Ethereum, Litecoin, and Monero. Several methods, i.e., HAR, ARFIMA, GARCH, LASSO, ridge regression, SVR, MLP, fuzzy neighbourhood model, random forest, and LSTM, are compared in terms of their forecasting accuracy. The realized variance calculated from intraday returns is used as the input variable for the models. Our experimental results demonstrate that there is no single best method for forecasting volatility of each cryptocurrency, and different models may perform better depending on the specific cryptocurrency, choice of the error metric and forecast horizon. Furthermore, we show that simple linear models such as HAR and ridge regression, perform not worse than more complex models like LSTM and random forest. The research provides a useful reference point for the development of more complex models and suggests potential benefits of incorporating additional input variables.

Keywords: Machine Learning, Cryptocurrency, Bitcoin, volatility, neural network