Home > Engineering > Computer Engineering > Special Issue > Smart Innovations in Computer Science and Applications > Beyond GARCH: A Comparative Study of Deep Learning Models (LSTM, GRU) vs Traditional Time Series Models for Cryptocurrency Volatility Forecasting

Beyond GARCH: A Comparative Study of Deep Learning Models (LSTM, GRU) vs Traditional Time Series Models for Cryptocurrency Volatility Forecasting

Call for Papers

Volume-10 | Issue-5

Last date : 27-Oct-2026

Best International Journal
Open Access | Peer Reviewed | Best International Journal | Indexing & IF | 24*7 Support | Dedicated Qualified Team | Rapid Publication Process | International Editor, Reviewer Board | Attractive User Interface with Easy Navigation

Journal Type : Open Access

First Update : Within 7 Days after submittion

Submit Paper Online

For Author

Research Area


Beyond GARCH: A Comparative Study of Deep Learning Models (LSTM, GRU) vs Traditional Time Series Models for Cryptocurrency Volatility Forecasting


Mayank Vijay Singh Kashyap



Mayank Vijay Singh Kashyap "Beyond GARCH: A Comparative Study of Deep Learning Models (LSTM, GRU) vs Traditional Time Series Models for Cryptocurrency Volatility Forecasting" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Special Issue | Smart Innovations in Computer Science and Applications, March 2026, pp.1557-1569, URL: https://www.ijtsrd.com/papers/ijtsrd102056.pdf

Because of the very unstable and unpredictable character of bitcoin assets, forecasting market volatility remains a difficult challenge for investors and risk managers. Traditional econometric models, such as GARCH, have been widely utilised for volatility forecasting, but they typically fail to capture the complex nonlinear patterns and abrupt market movements that occur with cryptocurrencies. Deep learning approaches have grown in prominence in recent years as a result of their improved capacity to handle complicated data patterns. However, there is still a paucity of research that properly compares the forecasting effectiveness of deep learning models to older approaches while also taking into account computing efficiency. The GARCH (1,1), EGARCH, and TGARCH models, together with deep learning architectures like Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU), are compared in this study in an effort to bridge this gap. From January 2020 to December 2023, the study uses a walk-forward validation technique to analyse daily Bitcoin and Ethereum price data in order to arrive at a trustworthy and accurate assessment. The findings reveal that deep learning models beat classic econometric techniques in prediction accuracy. The LSTM model decreases RMSE by roughly 18.2%, while the GRU model produces an MAE that is approximately 22.4% lower than the best-performing GARCH model. Traditional models, on the other hand, have a significant computational efficiency advantage since they require almost 200 times less training time than deep learning approaches. Statistical testing demonstrate that the performance differences are very significant at the 1% confidence level. Overall, this study provides researchers and practitioners with practical advice on how to select appropriate volatility prediction models based on their accuracy, interpretability, and computational resource requirements, as well as a clear benchmark comparison of traditional and contemporary forecasting techniques.

include deep learning models cryptocurrency, bitcoin, ethereum, GARCH models, long short-term memory (LSTM), volatility predictions, and comparative analysis.


IJTSRD102056
Special Issue | Smart Innovations in Computer Science and Applications, March 2026
1557-1569
IJTSRD | www.ijtsrd.com | E-ISSN 2456-6470
Copyright © 2019 by author(s) and International Journal of Trend in Scientific Research and Development Journal. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0) (http://creativecommons.org/licenses/by/4.0)

International Journal of Trend in Scientific Research and Development - IJTSRD having online ISSN 2456-6470. IJTSRD is a leading Open Access, Peer-Reviewed International Journal which provides rapid publication of your research articles and aims to promote the theory and practice along with knowledge sharing between researchers, developers, engineers, students, and practitioners working in and around the world in many areas like Sciences, Technology, Innovation, Engineering, Agriculture, Management and many more and it is recommended by all Universities, review articles and short communications in all subjects. IJTSRD running an International Journal who are proving quality publication of peer reviewed and refereed international journals from diverse fields that emphasizes new research, development and their applications. IJTSRD provides an online access to exchange your research work, technical notes & surveying results among professionals throughout the world in e-journals. IJTSRD is a fastest growing and dynamic professional organization. The aim of this organization is to provide access not only to world class research resources, but through its professionals aim to bring in a significant transformation in the real of open access journals and online publishing.

Thomson Reuters
Google Scholer
Academia.edu

ResearchBib
Scribd.com
archive

PdfSR
issuu
Slideshare

WorldJournalAlerts
Twitter
Linkedin