Please use this identifier to cite or link to this item:
https://hdl.handle.net/20.500.14279/2847
Title: | Overtraining in fuzzy ARTMAP: myth or reality? |
Authors: | Kasparis, Takis Georgiopoulos, Michael N. Koufakou, Anna |
metadata.dc.contributor.other: | Κασπαρής, Τάκης |
Major Field of Science: | Engineering and Technology |
Field Category: | Electrical Engineering - Electronic Engineering - Information Engineering |
Keywords: | Neural networks;Computational complexity;Fuzzy sets |
Issue Date: | Jul-2001 |
Source: | International Joint Conference on Neural Networks, 2001,Washington |
Conference: | IEEE International Conference on Neural Networks |
Abstract: | We examine the issue of overtraining in fuzzy ARTMAP. Over-training in fuzzy ARTMAP manifests itself in two different ways: 1) it degrades the generalization performance of fuzzy ARTMAP as training progresses; and 2) it creates unnecessarily large fuzzy ARTMAP neural network architectures. In this work we demonstrate that overtraining happens in fuzzy ARTMAP and propose an old remedy for its cure: cross-validation. In our experiments we compare the performance of fuzzy ARTMAP that is trained: 1) until the completion of training, 2) for one epoch, and 3) until its performance on a validation set is maximized. The experiments were performed on artificial and real databases. The conclusion derived from these experiments is that cross-validation is a useful procedure in fuzzy ARTMAP, because it produces smaller fuzzy ARTMAP architectures with improved generalization performance. The trade-off is that cross-validation introduces additional computational complexity in the training phase of fuzzy ARTMAP |
URI: | https://hdl.handle.net/20.500.14279/2847 |
ISSN: | 1098-7576 |
DOI: | 10.1109/IJCNN.2001.939529 |
Rights: | © 2001 IEEE |
Type: | Conference Papers |
Affiliation: | University of Central Florida |
Affiliation : | University of Central Florida |
Publication Type: | Peer Reviewed |
Appears in Collections: | Δημοσιεύσεις σε συνέδρια /Conference papers or poster or presentation |
CORE Recommender
Sorry the service is unavailable at the moment. Please try again later.
Items in KTISIS are protected by copyright, with all rights reserved, unless otherwise indicated.