Title :
Asymptotic second-order properties of sample partial correlations
Author_Institution :
Dept. of Autom. Control, Polytech. Inst. of Bucharest, Romania
fDate :
6/1/1989 12:00:00 AM
Abstract :
The asymptotic behavior of the sample partial correlations (PARCORs) is studied in the case of a general stationary process. An explicit formula for PARCOR´s second-order moments is provided. The analysis presented extends to more general processes as shown using previously reported results on the statistics of sample partial correlations of autoregressive (AR) and mixed autoregressive moving-average (ARMA) processes. This extension should be useful as the processes encountered in applications are not exactly AR or ARMA
Keywords :
correlation theory; matrix algebra; signal processing; statistical analysis; time series; AR processes; ARMA; PARCOR; asymptotic behavior; autoregressive moving-average; general stationary process; sample partial correlations; second-order moments; statistics; Acoustic signal processing; Direction of arrival estimation; Fault detection; Position measurement; Signal processing; Smoothing methods; Speech recognition; Statistical analysis; Sun; Surveillance;
Journal_Title :
Acoustics, Speech and Signal Processing, IEEE Transactions on
DOI :
10.1109/ASSP.1989.28069