Overdetermined spectral analysis of short data records

Dendrinos, M; Rangoussi, M; Carayannis, G

dc.contributor.author	Dendrinos, M	en
dc.contributor.author	Rangoussi, M	en
dc.contributor.author	Carayannis, G	en
dc.date.accessioned	2014-03-01T02:40:53Z
dc.date.available	2014-03-01T02:40:53Z
dc.date.issued	1988	en
dc.identifier.issn	02714310	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/30260
dc.subject	Signal To Noise Ratio	en
dc.subject	Spectral Analysis	en
dc.subject.other	Spectrum Analysis	en
dc.subject.other	Autocorrelation Estimators	en
dc.subject.other	Autocorrelation Matrix	en
dc.subject.other	Overdetermined Spectral Analysis	en
dc.subject.other	Parameters Hypersensitivity	en
dc.subject.other	Short Data Records	en
dc.subject.other	Signal Filtering and Prediction	en
dc.title	Overdetermined spectral analysis of short data records	en
heal.type	conferenceItem	en
heal.identifier.primary	10.1109/ISCAS.1988.15235	en
heal.identifier.secondary	http://dx.doi.org/10.1109/ISCAS.1988.15235	en
heal.publicationDate	1988	en
heal.abstract	The concept of using more than the minimal set of the normal equations was introduced in spectral analysis in order to counteract the hypersensitvity of the model's parameters. A simultaneous vertical and horizontal extension of the autocorrelation matrix leads to an overdetermined, order-extended system. The application of the overdetermined approach to the analysis of short signal frames is examined. A comparative study of the autocorrelation lags used in the overdetermined matrices, based on the biased and unbiased autocorrelation estimators, as well as their estimation performance, is presented. The authors suggest the use of the biased measure in the case of low SNR (signal-to-noise ratio) and short data records, since experimental results illustrate a sensitivity of the unbiased measure to the noise ratio and the frame length. However, when the conditions become better (higher SNR or more samples available) the unbiased measure tends to yield more accurate estimates than those of the biased one.	en
heal.publisher	Publ by IEEE, Piscataway, NJ, United States	en
heal.journalName	Proceedings - IEEE International Symposium on Circuits and Systems	en
dc.identifier.doi	10.1109/ISCAS.1988.15235	en
dc.identifier.volume	2	en
dc.identifier.spage	1585	en
dc.identifier.epage	1588	en