FORECASTING CRYPTOCURRENCY VOLUMES WITH ARTIFICIAL NEURAL NETWORKS AND SUPPORT VECTOR MACHINES
- 28 March 2026
- journal article
- Published by Journal of Empirical Economics and Social Sciences, Bandirma Onyedi Eylul University in Journal of Empirical Economics and Social Sciences
- Vol. 8 (1) , 27-46
Abstract
In this study, 50-day volume forecasts for Bitcoin, Ethereum, Binance Coin, and Ripple were conducted using Artificial Neural Networks and Support Vector Machines. The root mean square error, mean absolute error, and error autocorrelation at lag 1 were used to compare model performance. Analyses were carried out in R using data from investing.com. Findings indicate that the ANN model provides more accurate predictions for cryptocurrency volumes. According to 50-day forecasts, BTC volume is expected to increase, while BNB volume decreases with the ANN model but increases with the SVM model. For ETH and XRP, the ANN model indicates stable horizontal movements, whereas the SVM model predicts a sharp increase in ETH volume and a decline in XRP volume. Overall, although cryptocurrencies are innovative financial assets, their high volatility poses significant risks, suggesting they are not yet reliable investment instruments.Keywords
Funding Information
- No financial support was received.
This publication has 18 references indexed in Scilit:
- Intelligent forecasting with machine learning trading systems in chaotic intraday Bitcoin marketChaos, Solitons & Fractals, 2020
- Bitcoin price prediction using machine learning: An approach to sample dimension engineeringJournal of Computational and Applied Mathematics, 2020
- Volatility in the Cryptocurrency MarketOpen Economies Review, 2019
- Predicting the direction, maximum, minimum and closing prices of daily Bitcoin exchange rate using machine learning techniquesApplied Soft Computing, 2019
- Majority is not enoughCommunications of the ACM, 2018
- Bitcoin, gold and the dollar – A GARCH volatility analysisFinance Research Letters, 2016
- Efficient Learning Machines2015
- Double-spending fast payments in bitcoin2012
- Predicting good probabilities with supervised learning2005
- I.—COMPUTING MACHINERY AND INTELLIGENCEMind, 1950