Please use this identifier to cite or link to this item:
Title: Reproducing kernel hilbert space methods for cad tools
Authors: Chumachenko, S. V.
Khawar, Parvez
Gowher, Malik
Keywords: Reproducing Kernel Hilbert Space
Series Summation Method
Issue Date: 2004
Publisher: EWDTW
Citation: Chumachenko S.V. reproducing kernel hilbert space methods for cad tools/Chumachenko S.V., Khawar Parvez, Gowher Malik //Proceedings of East-West Design & Test Workshop (EWDTW’04)
Abstract: The review of known RKHS-methods for analysis of current state in science investigations is represented. The place of Series Summation Method in Reproducing Kernel Hilbert Space (RKHS) is determined. The new results obtained by this method are discussed. Reproducing Kernel Hilbert Space (RKHS) methods are interesting both pure theoretically and applied. RKHS theory has been a well studied topic, stemming from the original works of [1] to more recent studies on their application by [2, 3, 8-11]. Mathematical models based on RKHS and causal operators are presented in [3]. They are used at Pattern Recognition [4], Digital Data Processing [5], Image Compression [6], Computer Graphics [7]. Mentioned directions are described by mathematical tool – theory of wavelets [4]. RKHS methods are base tool in exact incremental learning [8], in statistical learning theory [2, 9]. The general theory of reproducing kernels which is combined with linear mappings in the framework of Hilbert spaces is considered in [2]. A framework for discussing the generalization ability of a trained network in the original function space using tools of functional analysis based on RKHS is introduced in [8]. Special kind of kernel based approximation scheme is also closely linked to regularization theory [10] and Support Vector Machines based approximation schemes [11] (Fig.).
Appears in Collections:Кафедра автоматизації проектування обчислювальної техніки (АПОТ)

Files in This Item:
File Description SizeFormat 
Чумаченко_EWDTW_2004.pdf686.11 kBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.