| Full text | |
| Author(s): |
Maciel, Leandro
;
Ballini, Rosangela
;
Gomide, Fernando
Total Authors: 3
|
| Document type: | Journal article |
| Source: | 2024 IEEE SYMPOSIUM ON COMPUTATIONAL INTELLIGENCE FOR FINANCIAL ENGINEERING AND ECONOMICS, CIFER 2024; v. N/A, p. 7-pg., 2024-01-01. |
| Abstract | |
Financial time series are complex, time-varying, noisy, and affected by many sources of uncertainty, making their forecasting a challenging task. This paper focuses on adaptive fuzzy modeling and forecasting of financial time series, emphasizing the evolving fuzzy modeling and the adaptive level set fuzzy modeling, two of the most prominent approaches of the current state of the art in adaptive fuzzy modeling. Adaptive level set fuzzy modeling (ALSM) uses the concept of level sets in a general data-driven adaptive framework. In ALSM, outputs are computed using functions that map the activation levels of the inputs into values in the output space. The method provides one-step-ahead predictions of price series from different markets, such as stocks, exchange rates, energy commodities, and cryptocurrencies. The results suggest that the fuzzy model can produce accurate forecasts of financial time series, especially when used to predict the future direction of prices, outperforming econometric time series methods, neural networks, and fuzzy-based approaches. (AU) | |
| FAPESP's process: | 20/09838-0 - BI0S - Brazilian Institute of Data Science |
| Grantee: | João Marcos Travassos Romano |
| Support Opportunities: | Research Grants - Research Centers in Engineering Program |