Annealed Competition of Experts for a Segmentation and Classification of Switching Dynamics

Klaus Pawelzik, Jens Kohlmorgen, Klaus Muller

Research output: Contribution to journalArticle

62 Citations (Scopus)

Abstract

We present a method for the unsupervised segmentation of data streams originating from different unknown sources that alternate in time. We use an architecture consisting of competing neural networks. Memory is included to resolve ambiguities of input-output relations. To obtain maximal specialization, the competition is adiabatically increased during training. Our method achieves almost perfect identification and segmentation in the case of switching chaotic dynamics where input manifolds overlap and input-output relations are ambiguous. Only a small dataset is needed for the training procedure. Applications to time series from complex systems demonstrate the potential relevance of our approach for time series analysis and short-term prediction.

Original languageEnglish
Pages (from-to)340-356
Number of pages17
JournalNeural Computation
Volume8
Issue number2
Publication statusPublished - 1996 Feb 15
Externally publishedYes

Fingerprint

Time series analysis
Large scale systems
Time series
Neural networks
Data storage equipment
Segmentation
Overlap
Time Series Analysis
Prediction
Complex Systems
Neural Networks

ASJC Scopus subject areas

  • Artificial Intelligence
  • Control and Systems Engineering
  • Neuroscience(all)

Cite this

Annealed Competition of Experts for a Segmentation and Classification of Switching Dynamics. / Pawelzik, Klaus; Kohlmorgen, Jens; Muller, Klaus.

In: Neural Computation, Vol. 8, No. 2, 15.02.1996, p. 340-356.

Research output: Contribution to journalArticle

Pawelzik, K, Kohlmorgen, J & Muller, K 1996, 'Annealed Competition of Experts for a Segmentation and Classification of Switching Dynamics', Neural Computation, vol. 8, no. 2, pp. 340-356.
Pawelzik, Klaus ; Kohlmorgen, Jens ; Muller, Klaus. / Annealed Competition of Experts for a Segmentation and Classification of Switching Dynamics. In: Neural Computation. 1996 ; Vol. 8, No. 2. pp. 340-356.
@article{7fe73c629e204044b357c22053050d08,
title = "Annealed Competition of Experts for a Segmentation and Classification of Switching Dynamics",
abstract = "We present a method for the unsupervised segmentation of data streams originating from different unknown sources that alternate in time. We use an architecture consisting of competing neural networks. Memory is included to resolve ambiguities of input-output relations. To obtain maximal specialization, the competition is adiabatically increased during training. Our method achieves almost perfect identification and segmentation in the case of switching chaotic dynamics where input manifolds overlap and input-output relations are ambiguous. Only a small dataset is needed for the training procedure. Applications to time series from complex systems demonstrate the potential relevance of our approach for time series analysis and short-term prediction.",
author = "Klaus Pawelzik and Jens Kohlmorgen and Klaus Muller",
year = "1996",
month = "2",
day = "15",
language = "English",
volume = "8",
pages = "340--356",
journal = "Neural Computation",
issn = "0899-7667",
publisher = "MIT Press Journals",
number = "2",

}

TY - JOUR

T1 - Annealed Competition of Experts for a Segmentation and Classification of Switching Dynamics

AU - Pawelzik, Klaus

AU - Kohlmorgen, Jens

AU - Muller, Klaus

PY - 1996/2/15

Y1 - 1996/2/15

N2 - We present a method for the unsupervised segmentation of data streams originating from different unknown sources that alternate in time. We use an architecture consisting of competing neural networks. Memory is included to resolve ambiguities of input-output relations. To obtain maximal specialization, the competition is adiabatically increased during training. Our method achieves almost perfect identification and segmentation in the case of switching chaotic dynamics where input manifolds overlap and input-output relations are ambiguous. Only a small dataset is needed for the training procedure. Applications to time series from complex systems demonstrate the potential relevance of our approach for time series analysis and short-term prediction.

AB - We present a method for the unsupervised segmentation of data streams originating from different unknown sources that alternate in time. We use an architecture consisting of competing neural networks. Memory is included to resolve ambiguities of input-output relations. To obtain maximal specialization, the competition is adiabatically increased during training. Our method achieves almost perfect identification and segmentation in the case of switching chaotic dynamics where input manifolds overlap and input-output relations are ambiguous. Only a small dataset is needed for the training procedure. Applications to time series from complex systems demonstrate the potential relevance of our approach for time series analysis and short-term prediction.

UR - http://www.scopus.com/inward/record.url?scp=0001558731&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0001558731&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0001558731

VL - 8

SP - 340

EP - 356

JO - Neural Computation

JF - Neural Computation

SN - 0899-7667

IS - 2

ER -