Physics
Scientific paper
Sep 1987
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=1987a%26a...183..156b&link_type=abstract
Astronomy and Astrophysics (ISSN 0004-6361), vol. 183, no. 1, Sept. 1987, p. 156-166. Research supported by D.A.N.O.F. and CNRS.
Physics
4
Astrolabes, Astrometry, Least Squares Method, Statistical Analysis, Data Reduction, Error Analysis, Linear Equations, Stellar Physics
Scientific paper
In this paper, the implicit assumptions under which the ordinary least-squares (OLS) fit is valid are discussed, with an application to the Paris-Astrolabe data. Nonzero mean of the computer residuals, serial dependence of the errors, heteroscedasticity, and nonnormality of the disturbance term in the model are also discussed. The suggested tests and remedies run into the problems of internal corrections, detection of omitted variables in the linear model such as sidereal time of passage, magnitude, and color effects, and into the problem of unequal variances of the errors in global reductions. Finally, the discussion of nonnormality of the 'statistical' residuals leads to the thorough study of the problem of the detection of suspicious stars and outliers in the OLS regressions, such as stars of the FK-4 supp. catalog which have systematically ill-behaved residuals in the program (Bougeard, 1986). This statistical analysis may be extended, through slight modifications, to many other astronomical data.
No associations
LandOfFree
Statistical problems about the use of the ordinary least-squares method in astrometry - Application to the Paris-Astrolabe data does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Statistical problems about the use of the ordinary least-squares method in astrometry - Application to the Paris-Astrolabe data, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Statistical problems about the use of the ordinary least-squares method in astrometry - Application to the Paris-Astrolabe data will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1667890