Backward adaptive biorthogonalization英文文献资料.docVIP

BACKWARDADAPTIVEBIORTHOGONALIZATION L.REBOLLO-NEIRA Received 7February 2003 A backward biorthogonalization approach is proposed, which modi?es biorthogonal func- tions so as to generate orthogonal projections onto a reduced subspace. The technique is relevant to problems amenable to be represented by a general linear model. In particular, problems ofdatacompression, noise reduction, andsparse representations maybetackled bytheproposed approach. 2000 Mathematics Subject Classi?cation: 46E99. 1. Introduction. We introduce a backward biorthogonalization technique relevant to the general linear model. Any data which may be described in terms of a linear combinationofwaveformssatis?esthismodel.Forinstance,theresponseofaphysical system toaparticular interaction varying asafunction ofaparameter, sayt,isoften represented byameasurable quantity f(t),which isamenable tobeexpressed inthe fashion N
f(t)= cnNαn(t), (1.1) n where the model waveforms αn, n = 1,...,N, are derivable by recourse to physical considerations. The determination of the coe?cients cN, n=1,...,N, entails solving n theinverse problem whenthefunction f(t) ismeasured. Thesuperscript N indicates that,unlessthewaveforms αn areorthogonal, theappropriate coe?cients cnN depend ontheorderofthemodel,thatis,thenumberN ofwaveforms αnbeingconsideredin (1.1). If such waveforms are linearly independent, then there exits a set of reciprocal functionsα?nN,n=1,...,N,whichisbiorthogonaltotheformer,thatis,
α?nN|αm=δn,m [5,11]. Here thesuperscript N indicates that thebiorthogonal functions α?N allow for n constructing orthogonal projections onto the subspace VN spanned by N waveforms αn.Hence,thecoe?cientscnN ofthelinearexpansion(1.1)approximating afunctionf atbest(inaminimumdistancesense)canbeobtainedbycomputingtheinnerproducts cnN=
α?nN|f[5,7]. Sincethereciprocal setα?nN,n=1,...,N (andtherefore thecoe?cients cnN),depends onthenumberN oftotalwaveforms, theyshouldberecalculated whenthisnumbe