247

15 ^{A Study and}

Experimentation on

Bitcoin Price Forecasting

with Time Series Analysis

and Recurrent Neural

Network (RNN)

Martin Aruldoss, V. Adarsh and

Miranda Lakshmi Travis

15.1  INTRODUCTION

Bitcoin is considered a widely recognized cryptocurrency in the digital world and is

one of the well-recognized investment assets (Chen et al., 2020c). Cryptocurrency

provides a platform to transfer currency from one end to another end safely without

any identification (anonymously). It has created an impact on the global economic

system, and this is not managed by any government or any centralized agencies. It

is mobilized digitally through cryptography and peer-to-peer protocols (Jalali and

Heidari, 2020). Many investors started to think about bitcoin as one of the very saf­

est investments compared with gold, platinum, safety bonds and so on. However,

during the pandemic situation resulting from the first and second waves of COVID-

19, the assumptions about bitcoin being the safest asset failed because of its drastic

fluctuations. Also, many queries have been posted on Google about the safety of this

decentralized investment (Chen et al., 2020a).

CONTENTS

15.1 Introduction...................................................................................................247

15.2 Literature Review..........................................................................................248

15.3 ARIMA with Recurrent Neural Network for Forecasting............................249

15.4 Experimentation............................................................................................250

15.5 Results and Discussion.................................................................................. 251

15.6 Conclusion.....................................................................................................255

References...............................................................................................................256

DOI:  10.1201/9781003138082-15