Convergence properties of a class of learning vector quantization algorithms
Authors:
- 1 February 1996
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 5 (2) , 361-368
Abstract
A mathematical analysis of a class of learning vector quantization (LVQ) algorithms is presented. Using an appropriate time-coordinate transformation, we show that the LVQ algorithms under consideration can be transformed into linear time-varying stochastic difference equations. Using this fact, we apply stochastic Lyapunov stability arguments, and we prove that the LVQ algorithms under consideration do indeed converge, provided that some appropriate conditions hold.Keywords
This publication has 18 references indexed in Scilit:
- Topologically ordered competitive samplingNeural Networks, 1994
- Generalized clustering networks and Kohonen's self-organizing schemeIEEE Transactions on Neural Networks, 1993
- 'Neural-gas' network for vector quantization and its application to time-series predictionIEEE Transactions on Neural Networks, 1993
- Analysis of the convergence properties of topology preserving neural networksIEEE Transactions on Neural Networks, 1993
- Competitive learning with generalized winner-take-all activationIEEE Transactions on Neural Networks, 1992
- A convergence theorem for Grossberg learningNeural Networks, 1990
- Derivation of a class of training algorithmsIEEE Transactions on Neural Networks, 1990
- Self-Organization and Associative Memory1988
- A stochastic model of retinotopy: A self organizing processBiological Cybernetics, 1986
- An Algorithm for Vector Quantizer DesignIEEE Transactions on Communications, 1980