Single layer linear feedforward neural network for signal estimation in the wavelet domain

Stephanakis Ioannis, M; Kollias, Stefanos

dc.contributor.author	Stephanakis Ioannis, M	en
dc.contributor.author	Kollias, Stefanos	en
dc.date.accessioned	2014-03-01T02:41:18Z
dc.date.available	2014-03-01T02:41:18Z
dc.date.issued	1997	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/30449
dc.subject	Approximation Method	en
dc.subject	Exponential Decay	en
dc.subject	Learning Algorithm	en
dc.subject	Rate of Convergence	en
dc.subject	Feedforward Neural Network	en
dc.subject	Low Frequency	en
dc.subject	Wavelet Transform	en
dc.subject.other	Approximation theory	en
dc.subject.other	Feedforward neural networks	en
dc.subject.other	Learning algorithms	en
dc.subject.other	Mathematical models	en
dc.subject.other	Wavelet transforms	en
dc.subject.other	Adaptive signal estimation model	en
dc.subject.other	Digital signal processing	en
dc.title	Single layer linear feedforward neural network for signal estimation in the wavelet domain	en
heal.type	conferenceItem	en
heal.identifier.primary	10.1109/ICDSP.1997.628382	en
heal.identifier.secondary	http://dx.doi.org/10.1109/ICDSP.1997.628382	en
heal.publicationDate	1997	en
heal.abstract	A model of adaptive signal estimation in the presence of noise is proposed in this paper using the wavelet transform and single layer linear feedforward neural networks. The networks successively estimate the residual of the approximations (details) of the input at different scales from coarse scales (low frequencies) to fine scales by minimizing the error of each channel of the wavelet transform separately. Since learning takes place in subspaces of the signal space one needs to train networks with fewer weights. The rate of convergence of the learning algorithm in several transform channels is investigated. It is shown that learning in the input space may diverge, whereas convergence is achieved in wavelet spaces. Experimental results are presented which compare the proposed approximation method with adaptive approximation using sigmoidal functions and exponentially decaying kernels.	en
heal.publisher	IEEE, Piscataway, NJ, United States	en
heal.journalName	International Conference on Digital Signal Processing, DSP	en
dc.identifier.doi	10.1109/ICDSP.1997.628382	en
dc.identifier.volume	2	en
dc.identifier.spage	487	en
dc.identifier.epage	490	en