A neural approach to data compression and classification
Recently, neural networks have evolved as an alternate approach instead of rule-based systems for data compression and automated solution of interpolation or classification problems. The most prominent feature of the neural processing paradigm is its inherent adaptability permitting fairly easy modification of a neural system to perform in a wide range of application environments. This paper presents the cosine classifier, a neural network model designed for unsupervised adaptation and solution of classification problems. Classification of hand-written digits is used to demonstrate its performance.
Unable to display preview. Download preview PDF.
- Carpenter, G.A. und Grossberg, S.: Neural Dynamics of Category Learning and Recognition: Attention, Memory Consolidation, and Amnesia, in: Davis, J., Newburgh, R. und Wegmann, E. (eds.): Brain Structure, Learning, and Memory, AAAS Symposium Series 1986.Google Scholar
- Grossberg, S.: Competitive Learning: From Interactive Activation to Adaptive Resonance, Cognitive Science, November 1987.Google Scholar
- Hinton, G.E., McClelland, J.L. und Rumelhart, D.E.: Distributed Representations, in: .Google Scholar
- Kohonen, T.: Self-Organization and Associative Memory, Berlin 1987.Google Scholar
- Kratzer, K.P.: Neuronale Netze — Grundlagen und Anwendung, München 1990.Google Scholar
- Krßel, U., Franke, J. und Schürmann, J.: Polynomklassifikator versus Multilayer-Perzeptron, 12th DAGM Symposium, 1990.Google Scholar
- Lippman, R.P.: An Introduction to Computing with Neural Nets, IEEE ASSP Magazine, April 1987.Google Scholar
- Rumelhart, D.E. und McClelland, J.L.: Parallel Distributed Processing: Explorations in the Microstructure of Cognition, Cambridge MA 1986.Google Scholar