ARIMA-GARCH Based Time Series Analysis of Cryptocurrency Volatility

Husna Sarirah  Husin; Yap Kah  Yong; Yodi; Anbuselvan  Sangodiah; Steven Eu Kok  Seng; R  Lalitha

Authors

Husna Sarirah Husin School of Computer Science, Faculty of Innovation & Technology, Taylor’s University, MALAYSIA
Yap Kah Yong Wawasan Open University, MALAYSIA
Yodi Fakultas Komputer, Universitas Universal, INDONESIA
Anbuselvan Sangodiah School of Computer Science, Faculty of Innovation & Technology, Taylor’s University, MALAYSIA
Steven Eu Kok Seng School of Engineering, Faculty of Innovation & Technology, Taylor’s University, MALAYSIA
R Lalitha Department of Computer Science and Engineering, Rajalakshmi Institute of Technology Chennai, INDIA

Keywords:

Cryptocurrency, time series, volatility analysis, ARIMA, GARCH, seasonality analysis

Abstract

Cryptocurrencies are a new type of asset that is changing the game. They are decentralized, very volatile, and their prices change quickly. Investors, policymakers, and researchers all need to know how bitcoin markets work. This article examines the volatility patterns of significant cryptocurrencies by employing the Autoregressive Integrated Moving Average (ARIMA) in conjunction with the GARCH model to assess historical price data and predict future trends. We also do a full volatility analysis to show how unpredictable cryptocurrency markets are. We also do a seasonality analysis and calculate the Relative Strength Index (RSI) for the technical analysis indicators. The study uses the R programming environment to clean up data, model time series, and check performance indicators. Our results show that ARIMA models accurately capture the time-based relationships in bitcoin time series, which makes them good for long-term predictions.