Treffer: Regional models: A new approach for nonlinear system identification via clustering of the self-organizing map : Advances in Self-Organizing Maps

Title:
Regional models: A new approach for nonlinear system identification via clustering of the self-organizing map : Advances in Self-Organizing Maps
Authors:
SOUZA, Amauri H, BARRETO, Guilherme A, CORONA, Francesco
Source:
Neurocomputing (Amsterdam). 147:31-46
Publisher Information:
Amsterdam: Elsevier, 2015.
Publication Year:
2015
Physical Description:
print, 63 ref
Original Material:
INIST-CNRS
Subject Terms:
Cognition, Computer science, Informatique, Sciences exactes et technologie, Exact sciences and technology, Sciences et techniques communes, Sciences and techniques of general use, Mathematiques, Mathematics, Probabilités et statistiques, Probability and statistics, Statistiques, Statistics, Inférence linéaire, régression, Linear inference, regression, Sciences appliquees, Applied sciences, Informatique; automatique theorique; systemes, Computer science; control theory; systems, Intelligence artificielle, Artificial intelligence, Connexionnisme. Réseaux neuronaux, Connectionism. Neural networks, Ajustement modèle, Model matching, Ajustamiento modelo, Algorithme Kohonen, Kohonen algorithm, Algoritmo Kohonen, Algorithme apprentissage, Learning algorithm, Algoritmo aprendizaje, Amas, Cluster, Montón, Analyse régression, Regression analysis, Análisis regresión, Autoorganisation, Self organization, Autoorganización, Classification, Clasificación, Entrée sortie, Input output, Entrada salida, Estimation M, M estimation, Estimación M, Estimation statistique, Statistical estimation, Estimación estadística, Evaluation performance, Performance evaluation, Evaluación prestación, Exogène, Exogenous, Exógeno, Identification système, System identification, Identificación sistema, Modèle autorégressif exogène, ARX model, Modelo autoregressivo exógeno, Modèle autorégressif, Autoregressive model, Modelo autorregresivo, Modèle linéaire, Linear model, Modelo lineal, Modèle non linéaire, Non linear model, Modelo no lineal, Modèle régression, Regression model, Modelo regresión, Modélisation, Modeling, Modelización, Méthode globale locale, Global local method, Método global local, Méthode moindre carré, Least squares method, Método cuadrado menor, Observation aberrante, Outlier, Observación aberrante, Partition, Partición, Prototype, Prototipo, Régression linéaire, Linear regression, Regresión lineal, Régression non linéaire, Non linear regression, Regresión no lineal, Réseau neuronal, Neural network, Red neuronal, Système non linéaire, Non linear system, Sistema no lineal, Théorie locale, Local theory, Teoría local, Machine d'apprentissage extrême, Extreme learning machine, Máquina de Aprendizado Extremo, Modèle donnée, Data models, Modelo de datos, Réseau neuronal non bouclé, Feedforward neural nets, Red neural unidireccional, Global models, Local models, Outliers, Robust regression, Self-Organizing Maps
Document Type:
Fachzeitschrift Article
File Description:
text
Language:
English
Author Affiliations:
Federal University of Ceará, Department of Teleinformatics Engineering, Av. Mister Hull, SIN, Center of Technology, Campus of Pici, CP 6005, CEP 60455-760 Fortaleza, Ceará, Brazil
Aalto University, Department of Information and Computer Science, Konemiehentie 2, Espoo, Finland
ISSN:
0925-2312
Access URL:
http://pascal-francis.inist.fr/vibad/index.php?action=search&terms=28836730
Rights:
Copyright 2015 INIST-CNRS
CC BY 4.0
Sauf mention contraire ci-dessus, le contenu de cette notice bibliographique peut être utilisé dans le cadre d’une licence CC BY 4.0 Inist-CNRS / Unless otherwise stated above, the content of this bibliographic record may be used under a CC BY 4.0 licence by Inist-CNRS / A menos que se haya señalado antes, el contenido de este registro bibliográfico puede ser utilizado al amparo de una licencia CC BY 4.0 Inist-CNRS
Notes:
Computer science; theoretical automation; systems

Mathematics
Accession Number:
edscal.28836730
Database:
PASCAL Archive

Weitere Informationen

Global modelling consists in fitting a single regression model to the available data, using the whole set of input and output observations. On the other side of the spectrum stands the local modelling approach, in which the input space is segmented into several small partitions and a specialized regression model is fit to each partition. In this paper, we propose a novel approach, called Regional Models (RM), that stands in between the global and local modelling ones. The proposal extends the two-level clustering approach by Vesanto and Alhoniemi (2000 [1]) to regression problems, more specifically, to system identification. In this regard, we first partition the input space using the Self-Organizing Map (SOM), and then perform clustering over the prototypes of the trained SOM. Finally, regional regression models are built over the clusters (i.e. over the regions) of SOM prototypes, not over each SOM prototype as in local modelling. Under the proposed framework, we build regional linear and nonlinear regression models. For the linear case, we use autoregressive models with eXogenous (ARX) whose parameters are estimated using the ordinary least-squares (OLS) method. Regional nonlinear ARX (NARX) models are built using the Extreme Learning Machine network. Additionally, we develop robust variants of the proposed regional models through the use of M-estimation, a statistical framework for handling outliers, since the OLS is highly sensitive to them. Comprehensive performance evaluation of the proposed models on synthetic and real-world datasets is carried out and the results compared to those achieved by standard global and local models.