Visualization of neural-network gaps based on error analysis

Author

Kantardzic, Mehmed M. ; Aly, Alaaeldin A. ; Elmaghraby, Adel S.

Author_Institution

Dept. of Eng. Math. & Comput. Sci., Louisville Univ., KY, USA

Volume

10

Issue

2

fYear

1999

fDate

3/1/1999 12:00:00 AM

Firstpage

419

Lastpage

426

Abstract

Presents a methodology for detection of neural-network gaps (NNGs) based on error analysis and the visualization that is applicable to the n-dimensional I/O domain. The generalization problem in artificial neural networks (ANN) training is analyzed and the concept of NNGs is introduced. The NNGs are highly undesirable in ANN generalization and methods for detecting, analyzing, and eliminating them are necessary. Previous methods for NNG detection, based on two-dimensional (2-D) and three dimensional (3-D) visualization, were not applicable for ANNs with more than three inputs. Experiments demonstrate advantages of this new methodology, which allows better understanding of the NNG phenomena using a quantitative approach

Keywords

error analysis; generalisation (artificial intelligence); learning (artificial intelligence); neural nets; error analysis; generalization problem; n-dimensional I/O domain; neural-network gaps; quantitative approach; visualization; Artificial neural networks; Computer networks; Error analysis; Neural networks; Particle measurements; Performance evaluation; Supervised learning; Testing; Two dimensional displays; Visualization;

fLanguage

English

Journal_Title

Neural Networks, IEEE Transactions on

Publisher

ieee

ISSN

1045-9227

Type

jour

DOI

10.1109/72.750572

Filename

750572