Visualization of changes in process dynamics using self-organizing maps

Authors:
Ignacio Díaz;Abel A. Cuadrado;Alberto B. Diez;Manuel Domínguez;Juan J. Fuertes;Miguel A. Prada
Affiliations:
University of Oviedo, Area de Ingeniería de Sistemas y Automática, Gijón, Spain;University of Oviedo, Area de Ingeniería de Sistemas y Automática, Gijón, Spain;University of Oviedo, Area de Ingeniería de Sistemas y Automática, Gijón, Spain;Instituto de Automática y Fabricación, Universidad de León, Escuela de Ingenierías, León, Spain;Instituto de Automática y Fabricación, Universidad de León, Escuela de Ingenierías, León, Spain;Instituto de Automática y Fabricación, Universidad de León, Escuela de Ingenierías, León, Spain
Venue:
ICANN'10 Proceedings of the 20th international conference on Artificial neural networks: Part II
Year:
2010

Citing 5
Cited 0

Self-Organizing Maps

Self-Organizing Maps
Novelty detection: a review—part 1: statistical approaches

Signal Processing
Novelty detection: a review—part 2: neural network based approaches

Signal Processing
A new approach to exploratory analysis of system dynamics using SOM. Applications to industrial processes

Expert Systems with Applications: An International Journal
Identification and control of dynamical systems using the self-organizing map

IEEE Transactions on Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper a new method based on the self-organizing map (SOM) is proposed to track and identify changes in the dynamic behaviour of a physical process. In a first stage, a SOM is trained on a parameter space composed of the coefficients of local dynamic models estimated around different operating points of the process. On execution, new models estimated from process data are compared against the stored models in the SOM to yield residual models that contain relevant information about the changes in the process dynamics. This information can be efficiently represented using time-frequency visualizations, that reveal unseen patterns in the frequency response and hide those that can be explained by the model.