Authors:
Helge Hülsen
and
Sergej Fatikow
Affiliation:
Division of Microrobotics and Control Engineering, University of Oldenburg, Germany
Keyword(s):
Associative Networks, Self-Organisation, Self-Supervised Learning, Extrapolation.
Related
Ontology
Subjects/Areas/Topics:
Adaptive Signal Processing and Control
;
Artificial Intelligence
;
Artificial Intelligence and Decision Support Systems
;
Computational Intelligence
;
Computer Vision, Visualization and Computer Graphics
;
Enterprise Information Systems
;
Feature Extraction
;
Features Extraction
;
Hybrid Learning Systems
;
Image and Video Analysis
;
Informatics in Control, Automation and Robotics
;
Information-Based Models for Control
;
Intelligent Control Systems and Optimization
;
Neural Networks Based Control Systems
;
Nonlinear Signals and Systems
;
Signal Processing, Sensors, Systems Modeling and Control
;
Soft Computing
;
System Identification
;
Time Series and System Modeling
Abstract:
Besides their typical classification task, Self-Organizing Maps (SOM) can be used to approximate input-output relations. They provide an economic way of storing the essence of past data into input/output support vector pairs. In this paper the SOLIM algorithm (Self-Organising Locally Interpolating Map) is reviewed and an extrapolation method is introduced. This framework allows finding one inverse of a nonlinear many-to-one mapping by exploiting the inherent neighbourhood criteria of the SOM part. Simulations show that the performance of the mapping including the extrapolation is comparable to other algorithms.