CobART: Correlation Based Adaptive Resonance Theory

Authors:
Mustafa Yavas;Ferda Nur Alpaslan
Affiliations:
Department of Computer Engineering, Middle East Technical University, Ankara, Turkey;Department of Computer Engineering, Middle East Technical University, Ankara, Turkey
Venue:
MED '09 Proceedings of the 2009 17th Mediterranean Conference on Control and Automation
Year:
2009

Citing 0
Cited 2

Hierarchical behavior categorization using correlation based adaptive resonance theory

Neurocomputing
Simple and complex behavior learning using behavior hidden Markov model and CobART

Neurocomputing

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper introduces a new type of ART 2 network that performs satisfactory categorization for a domain where the patterns are constructed from consecutive analog inputs. The main contribution relies on the correlation analysis methods used for category-matching. The resulting network model is named as Correlation Based Adaptive Resonance Theory (CobART). Correlation waveform analysis and Euclidian distance methods are used to elicit correlation between the learned categories and the data fed to the network.