Молікевич Р.С., кандидат географічних наук, доцент, доцент кафедри географії та екології, Херсонський державний університет, м. Херсон, Україна

ORCID ID: 0000-0002-6577-503X
e-mail: molikevych@gmail.com

A Strategy for Building a Cryptocurrency Investment Portfolio Using Machine Learning Models

Abstract. Introduction. This study uses machine learning models and optimization methods to present a strategic model for building a cryptocurrency investment portfolio. Specifically, the machine learning model XGBoost (Gradient Boosting) was used to predict cryptocurrency prices and identify optimal market entry points. The model revealed high cryptocurrency volatility and significant dependence on BTC market fluctuations. Based on Markowitz’s model, an investment portfolio was constructed to achieve the best predicted risk-return ratio. The portfolio consisted of 15 cryptocurrencies with high market capitalization but varying levels of volatility and risk. The market entry point was set based on the most recent price data available during the analysis. Seven months after the simulated purchase, the portfolio grew by nearly 50%. These results confirm the effectiveness of the proposed analysis and the optimal level of diversification.

Purpose. The purpose of this study is to develop a methodology for constructing a cryptocurrency investment portfolio using machine learning models and to evaluate the effectiveness of the portfolio based on simulated market performance.

Results. The study used the XGBoost price forecasting model and identified significant correlations between cryptocurrency prices and BTC market dynamics. The Markowitz model generated a diversified cryptocurrency portfolio that showed significant growth under simulated trading conditions. The results confirm the viability of using machine learning and optimization methods for cryptocurrency portfolio management.

Conclusions. Combining machine learning models, such as XGBoost, with portfolio optimization methods, such as Markowitz’s model, provides a powerful tool for effective decision-making in cryptocurrency investing. This approach allows investors to more accurately predict market dynamics, identify optimal entry points, and construct portfolios that balance risk and return. The use of the Sharpe ratio further enhances portfolio performance evaluation by quantifying the trade-off between volatility and profitability. In addition, including cryptocurrencies with varying levels of volatility and capitalization ensures a well-diversified portfolio that can withstand market fluctuations. The study demonstrates that a machine learning-driven strategy, combined with robust financial optimization techniques, mitigates risk and captures significant growth opportunities in the highly volatile cryptocurrency market. These findings underscore the value of adopting data-driven methods for portfolio management, and the potential to improve investment outcomes in emerging digital asset markets.

Keywords: Cryptocurrencies; machine learning; XGBoost; Markowitz model; price forecasting; investment portfolio; volatility; portfolio optimization; Sharpe ratio.

References:

1. Oliynyk, V. M., Frolov, S. M., & Leshchenko, Yu. I. (2012). Some aspects of optimization of the portfolio of financial instruments. Marketing and Innovation Management, 1(1), 140–147. http://nbuv.gov.ua/UJRN/Mimi_2012_1_17.
2. Aljinović, Z., Marasović, B., & Šestanović, T. (2021). Cryptocurrency Portfolio Selection—A Multicriteria Approach. Mathematics, 9(14), 1677. https://doi.org/10.3390/math9141677.
3. Bariviera, A. F., Basgall, M. J., Hasperué, W., & Naiouf, M. (2017). Some stylized facts of the Bitcoin market. Physica A: Statistical Mechanics and Its Applications, 484, 82–90. https://doi.org/10.1016/j.physa.2017.04.159.
4. Belcastro, L., Carbone, D., Cosentino, C., Fabrizio Marozzo, & Trunfio, P. (2023). Enhancing Cryptocurrency Price Forecasting by Integrating Machine Learning with Social Media and Market Data. Algorithms, 16(12), 542–542. https://doi.org/10.3390/a16120542.
5. Ciaian, P., Rajcaniova, M., & Kancs, d’Artis. (2015). The Economics of BitCoin Price Formation. Applied Economics, 48(19), 1799–1815. https://doi.org/10.1080/00036846.2015.1109038.
6. CoinGecko. (2024). Cryptocurrency Prices, Charts, and Crypto Market Cap | CoinGecko. CoinGecko. https://www.coingecko.com.
7. Coinmarketcap. (2024). Bitcoin. CoinMarketCap. https://coinmarketcap.com/currencies/bitcoin/.
8. James, G., Witten, D., Hastie, T., & Tibshirani, R. (2013). An Introduction to Statistical Learning. In Springer Texts in Statistics. Springer New York. https://doi.org/10.1007/978-1-4614-7138-7.
9. Kuo Chuen, D. L., Guo, L., & Wang, Y. (2017). Cryptocurrency: A new investment opportunity? The Journal of Alternative Investments, 20(3), 16–40. https://doi.org/10.3905/jai.2018.20.3.016.
10. Liu, W. (2018). Portfolio diversification across cryptocurrencies. Finance Research Letters, 29. https://doi.org/10.1016/j.frl.2018.07.010.
11. Rosa, P. D., Felber, P., & Schiavoni, V. (2024). CryptoAnalytics: Cryptocoins price forecasting with machine learning techniques. SoftwareX, 26, 101663–101663. https://doi.org/10.1016/j.softx.2024.101663.
12. Tay, F. E. H., & Cao, L. (2001). Application of support vector machines in financial time series forecasting. Omega, 29(4), 309–317. https://doi.org/10.1016/s0305-0483(01)00026-3
13. Trimborn, S., & Härdle, W. K. (2018). CRIX an Index for cryptocurrencies. Journal of Empirical Finance, 49, 107–122. https://doi.org/10.1016/j.jempfin.2018.08.004.
14. Trimborn, S., Li, M., & HHrdle, W. K. (2017). Investing with Cryptocurrencies – A Liquidity Constrained Investment Approach. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.2999782.
15. Yi, Y. (2024). Research on cryptocurrency portfolio based on Markowitz model. Highlights in Business Economics and Management, 32, 237–254. https://doi.org/10.54097/6q50bk41.