DocumentCode
1979658
Title
Data compression methods with using neurolike GTM structures
Author
Polishchuk, Uliana ; Tkachenko, Pavlo ; Tkachenko, Roman ; Yurchak, Irina
Author_Institution
Lviv Polytech. Nat. Univ., Lviv
fYear
2009
fDate
22-24 April 2009
Firstpage
68
Lastpage
69
Abstract
Applying neural networks for data compression provides compression with very good generalization characteristics with management losses or even without any information losses. In the basis of applying neural networks for data compression - is its training without supervisor (self-training).
Keywords
data compression; neural nets; unsupervised learning; data compression method; information loss; management loss; neural network; neurolike GTM structure; self-training method; Data compression; Electronic mail; Entropy; Management training; Multi-layer neural network; Neural networks; Neurons; Principal component analysis; Quantization; Signal generators; Auto-associative neural network; GTM neurolike structures; Geometrical Transformation Machine; Reducing of the data dimension; Reducing of the data multiplicity; Training without supervisor (self-training);
fLanguage
English
Publisher
ieee
Conference_Titel
Perspective Technologies and Methods in MEMS Design, 2009. MEMSTECH 2009. 2009 5th International Conference on
Conference_Location
Zakarpattya
Print_ISBN
978-966-2191-06-6
Type
conf
Filename
5069709
Link To Document