Vol.6, No.1, February 2017. ISSN: 2217-8309 eISSN: 2217-8333
TEM Journal
TECHNOLOGY, EDUCATION, MANAGEMENT, INFORMATICS Association for Information Communication Technology Education and Science |
Forecasting TRY/USD Exchange Rate with Various Artificial Neural Network Models
Cagatay Bal, Serdar Demir
© 2017 Cagatay Bal, Serdar Demir, published by UIKTEN. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License. (CC BY-NC-ND 4.0)
Citation Information: TEM Journal. Volume 6, Issue 1, Pages 11-16, ISSN 2217-8309, DOI: 10.18421/TEM61-02, February 2017.
Abstract:
Exchange rate forecasting is one of the most common subjects among the forecasting problem field. Researchers and academicians from many different disciplines proposed various approaches for better exchange rate forecasting. In recent years, for solving the stated forecasting problem artificial neural networks have become successful tool to obtain solutions. Many different artificial neural networks have been used, developed and still developing for even better and trustable forecasts. In this study, TRY/USD exchange rate forecasting is modeled with different learning algorithms, activations functions and performance measures. Various Artificial Neural Network (ANN) models for better forecasting were investigated, compared and the obtained forecasting results interpreted respectively. The results of the application show that Variable Learning Rate Backpropagation learning algorithm with tan-sigmoid activation function has the best performance for TRY/USD exchange rate forecasting.
Keywords – Activations functions, artificial neural networks, Exchange rates, Forecasting, Learning algorithms, Performance measures, TRY/USD. |
----------------------------------------------------------------------------------------------------------- ----------------------------------------------------------------------------------------------------------- |