عنوان مقاله [English]
Time series forecasting is an active research area that has drawn considerable attention for applications in a variety of areas. With the time series approach to forecasting, historical observations of the same variable are analyzed to develop a model describing the underlying relationship. Then, the established model is used in order to extrapolate the time series into the future. Improving forecasting, especially accurate time series forecasting, is an important yet often difficult task facing decision makers in many areas.
Computational intelligence approaches, such as artificial neural networks (ANNs) and fuzzy logic, have gradually established themselves as popular tools for forecasting complicated financial markets. Fuzzy is one of the most
important soft computing tools, which can provide a powerful framework in order to cope with vague or ambiguous problems, and can express linguistic values and human subjective judgments of natural language.
Artificial neural networks are flexible computing frameworks and universal approximators that can be applied to a wide range of forecasting problems with a high degree of accuracy. The major advantage of neural networks is their flexible nonlinear modeling capability. With ANNs, there is no need to specify a particular model form. Rather, the model is adaptively formed based on the features presented in the data. This data-driven approach is suitable for many empirical data sets, where no theoretical guidance is available to suggest an appropriate data generating process. Despite the advantages cited for them, ANNs have weaknesses, one of the most important of which is their requirement of large amounts of data in order to yield accurate results. Both theoretical
and empirical findings have indicated that integration of different models can be an effective way of improving upon their predictive performance and also overcoming the limitations of single models, especially when the models in combination are quite different.
In this paper, a new hybrid model of artificial neural networks is proposed based on the basic concepts of fuzzy logic, in order to overcome the data restriction of neural networks and yield more accurate results than traditional
ANNs in situations of short time spans. In the proposed model, instead of using crisp parameters in each layer, fuzzy parameters in the form of triangular fuzzy numbers are applied for related parameters of these layers. In this way, the proposed model can search the feasible spaces easily and more efficiently for finding the optimum values of parameters. The empirical results of exchange rate forecasting indicate that the hybrid model is more satisfactory than its components, i.e, artificial neural networks and fuzzy regression models.