A class of optimization problems on minimizing variance based criteria in respect of parameter estimators of a linear model. Issue 10 (2nd October 2020)

• Record Type: Journal Article
• Authors: Chowdhury, M.; Chen, M.; Mandal, S.
• Is Part Of: Communications in statistics. Volume 49:Issue 10(2020)
• Page Start: 2719
• Page End: 2731
• Publication Date: 2020-10-02
• Subjects: Optimizing probability distributions -- Parameter estimation -- Optimal designs -- Directional derivatives -- Multiplicative algorithms; 62F10 -- 62K05; Mathematical statistics -- Periodicals; Mathematical statistics -- Data processing -- Periodicals; Digital computer simulation -- Periodicals; 519.5
• DOI: 10.1080/03610918.2018.1529240
• Languages: English
• ISSNs: 0361-0918

Abstract

Abstract: We first consider an optimization problem in which we minimize the average or total variance of the estimators of some parameters of interest in a linear model. In some models, subsets of parameters may be of more interest in this respect than all parameters. We then consider a second optimization problem in which we construct designs achieving equality of variances of the estimators of two linear functions of the parameters. The approaches are formulated for a general regression model, and are explored through some models of interest.