Author(s)

Meiyu Song, Yueqi Yu, Yile Zhang, Zizhen Wang

Corresponding Author

Meiyu Song

Abstract

Model predictions are very important for the cryptocurrency market, and accurate model predictions are important for the application of blockchain in the market and to improve the socio-economic benefits. In order to predict the realized fluctuation of cryptocurrency more accurately, based on the four models of BASELINE MODEL, RANDOM WALK, GARCH, NEURAL NETWORK, combining the current situation at home and abroad and the advantages in big data processing, by comparing the metrics of RMSE and RMSPE, we conclude that the optimal model for the realized fluctuation is random walk.[1]

Keywords

ETH, GARCH, neural network, forecasting