Managing extreme cryptocurrency volatility in algorithmic trading: EGARCH via genetic algorithms and neural networks
The blockchain ecosystem has seen a huge growth since 2009, with the introduction of Bitcoin, driven by conceptual and algorithmic innovations, along with the emergence of numerous new cryptocurrencies. While significant attention has been devoted to established cryptocurrencies like Bitcoin and Eth...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
AIMS Press
2024-03-01
|
| Series: | Quantitative Finance and Economics |
| Subjects: | |
| Online Access: | https://www.aimspress.com/article/doi/10.3934/QFE.2024007?viewType=HTML |
| _version_ | 1827278468647223296 |
|---|---|
| author | David Alaminos M. Belén Salas Ángela M. Callejón-Gil |
| author_facet | David Alaminos M. Belén Salas Ángela M. Callejón-Gil |
| author_sort | David Alaminos |
| collection | DOAJ |
| description | The blockchain ecosystem has seen a huge growth since 2009, with the introduction of Bitcoin, driven by conceptual and algorithmic innovations, along with the emergence of numerous new cryptocurrencies. While significant attention has been devoted to established cryptocurrencies like Bitcoin and Ethereum, the continuous introduction of new tokens requires a nuanced examination. In this article, we contribute a comparative analysis encompassing deep learning and quantum methods within neural networks and genetic algorithms, incorporating the innovative integration of EGARCH (Exponential Generalized Autoregressive Conditional Heteroscedasticity) into these methodologies. In this study, we evaluated how well Neural Networks and Genetic Algorithms predict "buy" or "sell" decisions for different cryptocurrencies, using F1 score, Precision, and Recall as key metrics. Our findings underscored the Adaptive Genetic Algorithm with Fuzzy Logic as the most accurate and precise within genetic algorithms. Furthermore, neural network methods, particularly the Quantum Neural Network, demonstrated noteworthy accuracy. Importantly, the X2Y2 cryptocurrency consistently attained the highest accuracy levels in both methodologies, emphasizing its predictive strength. Beyond aiding in the selection of optimal trading methodologies, we introduced the potential of EGARCH integration to enhance predictive capabilities, offering valuable insights for reducing risks associated with investing in nascent cryptocurrencies amidst limited historical market data. This research provides insights for investors, regulators, and developers in the cryptocurrency market. Investors can utilize accurate predictions to optimize investment decisions, regulators may consider implementing guidelines to ensure fairness, and developers play a pivotal role in refining neural network models for enhanced analysis. |
| first_indexed | 2024-04-24T07:59:24Z |
| format | Article |
| id | doaj.art-8518261123064444885930035f035e09 |
| institution | Directory Open Access Journal |
| issn | 2573-0134 |
| language | English |
| last_indexed | 2024-04-24T07:59:24Z |
| publishDate | 2024-03-01 |
| publisher | AIMS Press |
| record_format | Article |
| series | Quantitative Finance and Economics |
| spelling | doaj.art-8518261123064444885930035f035e092024-04-18T01:47:36ZengAIMS PressQuantitative Finance and Economics2573-01342024-03-018115320910.3934/QFE.2024007Managing extreme cryptocurrency volatility in algorithmic trading: EGARCH via genetic algorithms and neural networksDavid Alaminos0 M. Belén Salas 1Ángela M. Callejón-Gil21. Department of Business, University of Barcelona, Barcelona, Spain2. Department of Finance and Accounting, University of Málaga, Málaga, Spain 3. Cátedra de Economía y Finanzas Sostenibles, University of Málaga, Málaga, Spain2. Department of Finance and Accounting, University of Málaga, Málaga, Spain 3. Cátedra de Economía y Finanzas Sostenibles, University of Málaga, Málaga, SpainThe blockchain ecosystem has seen a huge growth since 2009, with the introduction of Bitcoin, driven by conceptual and algorithmic innovations, along with the emergence of numerous new cryptocurrencies. While significant attention has been devoted to established cryptocurrencies like Bitcoin and Ethereum, the continuous introduction of new tokens requires a nuanced examination. In this article, we contribute a comparative analysis encompassing deep learning and quantum methods within neural networks and genetic algorithms, incorporating the innovative integration of EGARCH (Exponential Generalized Autoregressive Conditional Heteroscedasticity) into these methodologies. In this study, we evaluated how well Neural Networks and Genetic Algorithms predict "buy" or "sell" decisions for different cryptocurrencies, using F1 score, Precision, and Recall as key metrics. Our findings underscored the Adaptive Genetic Algorithm with Fuzzy Logic as the most accurate and precise within genetic algorithms. Furthermore, neural network methods, particularly the Quantum Neural Network, demonstrated noteworthy accuracy. Importantly, the X2Y2 cryptocurrency consistently attained the highest accuracy levels in both methodologies, emphasizing its predictive strength. Beyond aiding in the selection of optimal trading methodologies, we introduced the potential of EGARCH integration to enhance predictive capabilities, offering valuable insights for reducing risks associated with investing in nascent cryptocurrencies amidst limited historical market data. This research provides insights for investors, regulators, and developers in the cryptocurrency market. Investors can utilize accurate predictions to optimize investment decisions, regulators may consider implementing guidelines to ensure fairness, and developers play a pivotal role in refining neural network models for enhanced analysis.https://www.aimspress.com/article/doi/10.3934/QFE.2024007?viewType=HTMLemerging cryptocurrenciesegarchgenetic algorithmsneural networksalgorithmic tradingquantum computingdeep learning |
| spellingShingle | David Alaminos M. Belén Salas Ángela M. Callejón-Gil Managing extreme cryptocurrency volatility in algorithmic trading: EGARCH via genetic algorithms and neural networks Quantitative Finance and Economics emerging cryptocurrencies egarch genetic algorithms neural networks algorithmic trading quantum computing deep learning |
| title | Managing extreme cryptocurrency volatility in algorithmic trading: EGARCH via genetic algorithms and neural networks |
| title_full | Managing extreme cryptocurrency volatility in algorithmic trading: EGARCH via genetic algorithms and neural networks |
| title_fullStr | Managing extreme cryptocurrency volatility in algorithmic trading: EGARCH via genetic algorithms and neural networks |
| title_full_unstemmed | Managing extreme cryptocurrency volatility in algorithmic trading: EGARCH via genetic algorithms and neural networks |
| title_short | Managing extreme cryptocurrency volatility in algorithmic trading: EGARCH via genetic algorithms and neural networks |
| title_sort | managing extreme cryptocurrency volatility in algorithmic trading egarch via genetic algorithms and neural networks |
| topic | emerging cryptocurrencies egarch genetic algorithms neural networks algorithmic trading quantum computing deep learning |
| url | https://www.aimspress.com/article/doi/10.3934/QFE.2024007?viewType=HTML |
| work_keys_str_mv | AT davidalaminos managingextremecryptocurrencyvolatilityinalgorithmictradingegarchviageneticalgorithmsandneuralnetworks AT mbelensalas managingextremecryptocurrencyvolatilityinalgorithmictradingegarchviageneticalgorithmsandneuralnetworks AT angelamcallejongil managingextremecryptocurrencyvolatilityinalgorithmictradingegarchviageneticalgorithmsandneuralnetworks |