Sample selection and training of self-organizing map neural network in multiple models approximation

Author

Gao, Dayuan ; Zhu, Hai ; Liu, Xijing ; Wang, Chao

Author_Institution

Dept. of Navig. & Commun., Navy Submarine Acad., Qingdao, China

fYear

2012

fDate

6-8 July 2012

Firstpage

3053

Lastpage

3058

Abstract

The self-organizing map (SOM) neural network has been used widely in multiple models approximation (MMA). However, the clustering property of SOM may not be fit for MMA. This paper introduces the idea of active learning into the training of SOM, especially for MMA. The neural network selects actively the training samples according to the approximation error of local models. As a result, the distribution of the neural nodes is changed so that the performance of MMA is improved. The process of this training method and the performance improvement are illustrated by a simulation example.

Keywords

approximation theory; learning (artificial intelligence); self-organising feature maps; MMA; SOM training; active learning; approximation error; clustering property; multiple models approximation; neural nodes distribution; self-organizing map neural network; Approximation error; Computational modeling; Data models; Neural networks; Training; Training data; Multiple Models Approximation; Neural Network; Self-Organizing Map;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Control and Automation (WCICA), 2012 10th World Congress on

Conference_Location

Beijing

Print_ISBN

978-1-4673-1397-1

Type

conf

DOI

10.1109/WCICA.2012.6358395

Filename

6358395