Abstract: The paper presents the experiment on the time series whose elements are month values of BIS effective exchange rate of USD from January 1994 till March 2010. A tendency of BIS (Bank of International Settlements) effective exchange rate to increase or decrease is an expected value. First, a process of building of the neural network for events forecasting is presented, that is the selection of networks’ architecture and parameters. Next, the effect of adding data calibrated by nonlinear input function to input data calibrated linearly is described. The nonlinear input function - hyperbolic tangent was accepted. Hyperbolic tangent sigmoid transfer function and log sigmoid transfer function are commonly used as transfer functions in neural networks.

Keywords: neural network, time series.

ACM Classification Keywords: I.2.8 Data calibration.

Link:

THE EFFECT OF INTRODUCTION OF THE NON-LINEAR CALIBRATION FUNCTION AT THE INPUT OF THE NEURAL NETWORK

Piotr Romanowski

http://www.foibg.com/ibs_isc/ibs-26/ibs-26-p19.pdf