Premium
On Estimators Obtained From a Sample Augmented by Multiple Regression
Author(s) -
Moran M. A.
Publication year - 1974
Publication title -
water resources research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.863
H-Index - 217
eISSN - 1944-7973
pISSN - 0043-1397
DOI - 10.1029/wr010i001p00081
Subject(s) - estimator , random variate , statistics , mathematics , variance (accounting) , sample (material) , control variates , regression , maximum likelihood , monte carlo method , random variable , hybrid monte carlo , chemistry , accounting , chromatography , markov chain monte carlo , business
A sample of N observations is taken from a p + 1 variate normal distribution. The first N observations include values on all p + 1 variates, whereas the remaining N ‐ n observations include values for p of the variates only. This paper reviews the properties of the estimators that use the N complete observations on the p variates to improve estimation of the mean and variance of the variate with only n observations. In particular, the relationship of suggested estimators to maximum likelihood estimators, corrected for bias, is given. The general advantages and limitations of such estimators are discussed.