Design of RBF network based on fuzzy clustering method for modeling of respiratory system

  • Authors:

  • Kouji Maeda;Shunshoku Kanae;Zi-Jiang Yang;Kiyoshi Wada

  • Affiliations:

  • Department of Electrical and Electronic Systems Engineering, Graduate School of Information Science and Electrical Engineering, Kyushu University, Email:jin@ees.kyushu-u.ac.jp, Fukuoka, Japan;Department of Electrical and Electronic Systems Engineering, Graduate School of Information Science and Electrical Engineering, Kyushu University, Email:jin@ees.kyushu-u.ac.jp, Fukuoka, Japan;Department of Electrical and Electronic Systems Engineering, Graduate School of Information Science and Electrical Engineering, Kyushu University, Email:jin@ees.kyushu-u.ac.jp, Fukuoka, Japan;Department of Electrical and Electronic Systems Engineering, Graduate School of Information Science and Electrical Engineering, Kyushu University, Email:jin@ees.kyushu-u.ac.jp, Fukuoka, Japan

  • Venue:
  • ISNN'06 Proceedings of the Third international conference on Advances in Neural Networks - Volume Part III
  • Year:
  • 2006

Quantified Score

Hi-index 0.01

Visualization

Abstract

Pulmonary elastance provides an important basis for deciding air pressure parameters of mechanical ventilators, and airway resistance is an important parameter in the diagnosis of respiratory diseases. The authors have proposed a second order nonlinear differential equation model of respiratory system whose elastic and resistant coefficients are expressed by RBF networks with the lung volume as the input. When we use RBF networks expression, numerical stability can be expected, because the output of each node is in range of [0,1], the balance between each node is good. However, the problems of deciding the number of nodes and the center/deviation of each node were remained. In this paper, a design method of RBF network based on fuzzy clustering method is proposed to decide center and deviation of each node. By means of fuzzy clustering, the available data set is partitioned into fuzzy subsets so that each RBF works effectively. The proposed method is validated by examples of application to practical clinical data.