Forecasting digital asset return: an application of machine learning model
Forecasting digital asset return: an application of machine learning model
In this study, we aim to identify the machine learning model that can overcome the limitations of traditional statistical modelling techniques in forecasting Bitcoin prices. Also, we outline the necessary conditions that make the model suitable. We draw on a multivariate large data set of Bitcoin prices and its market microstructure variables and apply three machine learning models, namely double deep Q-learning, XGBoost and ARFIMA-GARCH. The findings show that the double deep Q-learning model outperforms the others in terms of returns and Sortino ratio and is capable of one-step-ahead sign forecast of the returns even on synthetic data. These critical insights in forecasting literature will support practitioners and regulators to identify an economically viable cryptocurrency forecasting return model.
bitcoin, digital asset, double deep Q-learning, forecasting price, machine learning, reinforcement learning, time-series
Ciciretti, Vito
e6b12ccd-63bd-444d-986e-8cab021ce6f9
Pallotta, Alberto
f980cc20-2fe0-4f27-b6d9-e303e9b469b5
Lodh, Suman
0c3fe0ce-1de3-4d75-9957-9edce8b81a30
Senyo, P.K.
b2150f66-8ef9-48f7-af32-3b055d4fa691
Nandy, Monomita
bd32ea6b-7baf-45d2-9f86-953d0edc2045
18 November 2024
Ciciretti, Vito
e6b12ccd-63bd-444d-986e-8cab021ce6f9
Pallotta, Alberto
f980cc20-2fe0-4f27-b6d9-e303e9b469b5
Lodh, Suman
0c3fe0ce-1de3-4d75-9957-9edce8b81a30
Senyo, P.K.
b2150f66-8ef9-48f7-af32-3b055d4fa691
Nandy, Monomita
bd32ea6b-7baf-45d2-9f86-953d0edc2045
Ciciretti, Vito, Pallotta, Alberto, Lodh, Suman, Senyo, P.K. and Nandy, Monomita
(2024)
Forecasting digital asset return: an application of machine learning model.
International Journal of Finance & Economics.
(doi:10.1002/ijfe.3062).
Abstract
In this study, we aim to identify the machine learning model that can overcome the limitations of traditional statistical modelling techniques in forecasting Bitcoin prices. Also, we outline the necessary conditions that make the model suitable. We draw on a multivariate large data set of Bitcoin prices and its market microstructure variables and apply three machine learning models, namely double deep Q-learning, XGBoost and ARFIMA-GARCH. The findings show that the double deep Q-learning model outperforms the others in terms of returns and Sortino ratio and is capable of one-step-ahead sign forecast of the returns even on synthetic data. These critical insights in forecasting literature will support practitioners and regulators to identify an economically viable cryptocurrency forecasting return model.
Text
FullText
- Accepted Manuscript
Text
Int J Fin Econ - 2024 - Ciciretti - Forecasting Digital Asset Return An Application of Machine Learning Model
- Version of Record
More information
Accepted/In Press date: 12 October 2024
e-pub ahead of print date: 18 November 2024
Published date: 18 November 2024
Keywords:
bitcoin, digital asset, double deep Q-learning, forecasting price, machine learning, reinforcement learning, time-series
Identifiers
Local EPrints ID: 496078
URI: http://eprints.soton.ac.uk/id/eprint/496078
ISSN: 1076-9307
PURE UUID: 69c9221f-0747-497c-9cbc-00eaf2e7dce7
Catalogue record
Date deposited: 03 Dec 2024 17:35
Last modified: 06 Dec 2024 03:00
Export record
Altmetrics
Contributors
Author:
Vito Ciciretti
Author:
Alberto Pallotta
Author:
Suman Lodh
Author:
Monomita Nandy
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics
