Presentation 2001/10/12
A Study on Evolutionary Design of Neural Network Trees with Nodes of Limited Fan-in-fan-out Neural Networks
Shinichi MIZUNO, Qiangfu ZHAO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Neural network tree(NNTree)is a kind of models in the context of machine learning. By integrating symbolic approach and non-symbolic approach, NNTrees can achieve high recognition rate with comparatively less nodes than traditional decision trees. However, how each expert neural network(ENN), which is a node of NNTrees, is designed is still an important problem. In this paper, we study NNTrees with nodes of limited fan-in-fan-out and show several experimental results. In the experiments, we were able to impose the limitation on NNTrees without steep increase on the number of nodes. Such NNTrees are effective for extracting precise rules. In addition, we can guess which sub-pattern is important for pattern recognition.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) decision trees / neural networks / pattern recognition / machine learning
Paper # NC2001-59
Date of Issue

Conference Information
Committee NC
Conference Date 2001/10/12(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language ENG
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A Study on Evolutionary Design of Neural Network Trees with Nodes of Limited Fan-in-fan-out Neural Networks
Sub Title (in English)
Keyword(1) decision trees
Keyword(2) neural networks
Keyword(3) pattern recognition
Keyword(4) machine learning
1st Author's Name Shinichi MIZUNO
1st Author's Affiliation The University of Aizu Graduate School of Computer Science and Engineering()
2nd Author's Name Qiangfu ZHAO
2nd Author's Affiliation The University of Aizu Graduate School of Computer Science and Engineering
Date 2001/10/12
Paper # NC2001-59
Volume (vol) vol.101
Number (no) 365
Page pp.pp.-
#Pages 6
Date of Issue