A new criterion of NN structure selection for financial forecasting

Author

Perrone, Antonio L. ; Basti, Gianfranco

Author_Institution

Pontifical Lateran Univ., Rome, Italy

Volume

6

fYear

1999

fDate

36342

Firstpage

3898

Abstract

For the evaluation and the selection of the optimal neural net (NN) structure complexity, as a function of the minimization either of the approximation error or of the generalization error, we discuss briefly the minimum description length (MDL) method. Because of the theoretical and practical limitations of this criterion-overall for stochastic time series previsions-we shortly introduce our new dynamic sampling window (DSW) method for the optimal NN structure definition for financial forecasting

Keywords

finance; forecasting theory; minimisation; neural net architecture; time series; DSW method; MDL method; NN structure selection criterion; approximation error minimization; dynamic sampling window method; financial forecasting; generalization error minimization; minimum description length method; optimal neural net structure complexity; stochastic time series previsions; Approximation error; Cost function; Laboratories; Length measurement; Minimization methods; Neural networks; Phase estimation; Predictive models; Sampling methods; Stochastic processes;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 1999. IJCNN '99. International Joint Conference on

Conference_Location

Washington, DC

ISSN

1098-7576

Print_ISBN

0-7803-5529-6

Type

conf

DOI

10.1109/IJCNN.1999.830778

Filename

830778