Title :
The extended-least-squares treatment of correlated data
Author :
Cohen, E.R. ; Tuninsky, V.S.
Author_Institution :
Sci. Center, Rockwell Int. Corp., Thousand Oaks, CA, USA
fDate :
June 27 1994-July 1 1994
Abstract :
A generalization of the extended-least-squares algorithms for the case of correlated discrepant data is given. The expressions of the linear, unbiased, minimum-variance estimators (LUMVE) derived before are reformulated. A posteriori estimates of the variance taking into account the inconsistency of all of the experimental data have the same form as for the case of non-correlated data. These estimates extend the previous improvement on the "traditional" Birge-ratio procedures to the case of correlated input data.<>
Keywords :
correlation theory; data analysis; estimation theory; least squares approximations; statistical analysis; Birge-ratio; LUMVE; a posteriori estimates; correlated discrepant data; extended-least-squares algorithms; linear unbiased minimum-variance estimators; Covariance matrix; Equations; Metrology; Particle measurements; Probability distribution; Statistical analysis; Testing;
Conference_Titel :
Precision Electromagnetic Measurements, 1994. Digest., 1994 Conference on
Conference_Location :
Boulder, CO, USA
Print_ISBN :
0-7803-1984-2
DOI :
10.1109/CPEM.1994.333253
