Automatic Complexity Control for System Identification
- Ali Ghodsi, University of Waterloo
- Dale Schuurmans, AICML
As a prerequisite for system identi cation based on c-mean clustering (FCM), it is necessary to assign the number of underlying partitions to be used for a given data set. However, for the FCM clustering algorithm it is not known how to assign the number of clusters optimally a priori, and the problem of selecting an appropriate number of clusters is usually treated heuristically. In this paper we derive a theoretical criterion for assigning the appropriate number of clusters. We
Citation
A. Ghodsi, D. Schuurmans. "Automatic Complexity Control for System Identification". Fuzzy Systems Association World Congress(IFSA), June 2003.| Keywords: | |
| Category: | In Conference |
BibTeX
@incollection{Ghodsi+Schuurmans:IFSA03,
author = {Ali Ghodsi and Dale Schuurmans},
title = {Automatic Complexity Control for System Identification},
booktitle = {Fuzzy Systems Association World Congress(IFSA)},
year = 2003,
}Last Updated: June 01, 2007Submitted by Staurt H. Johnson
