A COMPARISON BETWEEN THREE PREDICTIVE MODELS OF COMPUTATIONAL INTELLIGENCE
Time series prediction is an open problem and many researchers are trying to find new predictive methods and improvements for the existing ones. Lately methods based on neural networks are used extensively for time series prediction. Also, support vector machines have solved some of the problems faced by neural networks and they began to be widely used for time series prediction. The main drawback of those two methods is that they are global models and in the case of a chaotic time series it is unlikely to find such model. In this paper it is presented a comparison between three predictive from computational intelligence field one based on neural networks one based on support vector machine and another based on chaos theory. We show that the model based on chaos theory is an alternative to the other two methods. Keywords: time series prediction, neural networks, support vector machines, chaos theory, exchange rate.
Year of publication: |
2013
|
---|---|
Authors: | CIOBANU, DUMITRU ; VIOLETA, BAR MARY |
Published in: |
Annals - Economy Series. - Facultatea de Ştiinţe Economice şi Gestiunea Afacerilor, ISSN 1844-7007. - Vol. 6.2013, December, p. 62-74
|
Publisher: |
Facultatea de Ştiinţe Economice şi Gestiunea Afacerilor |
Saved in:
Saved in favorites
Similar items by person
-
ON THE PREDICTION OF EXCHANGE RATE DOLLAR/EURO WITH AN SVM MODEL
Dumitru, CIOBANU, (2013)
-
A Comparison Between Two Predictive Models of Artificial Intelligence
Dumitru, Ciobanu, (2012)
-
Ciobanu, Dumitru, (2012)
- More ...