Third Annual Conference on Forecasting Financial Markets: New Advances for Exchange Rates, Interest Rates, and Asset Management, Londres (Reino Unido). 27-29 Marzo 1996
Resumen:
In this paper the analysis of the influence input variables have on the output of a fitted model is presented. It is based on the statistical study of the derivatives of the output of the model with regards to its standardized inputs (also called I/O sensitivities). We have tested the technique on the problem of forecasting the French T-Bond. In the second part we exploit these results to design a more reachable task, which is also more useful in the decision process of the traders. We finally obtain a decision making system which would be implemented in the trading room of the Caisse Nationales du Crédit Agricole (CNCA).
Palabras clave: Radial Basis Function Network, Recurrent networks, variable selection
Fecha de publicación: marzo 1996.
Cita:
Dorizzi, B., Pellieux, G., Jacquet, F., Czernichow, T., Muñoz, A., Selecting the relevant variables to forecast the Frech T-Bond, Third Annual Conference on Forecasting Financial Markets: New Advances for Exchange Rates, Interest Rates, and Asset Management, Londres (Reino Unido). 27-29 Marzo 1996.
IIT-96-027A