Maximum posterior estimate and its statistic properties

Abstract

Based on an extended Gauss-Markov model where the unknown parameters has the prior normal distribution, this paper derives the maximum posterior estimate formulas of the parameters which are proved to be unbiased, efficient, and of variance of unit weight which is biased. Finally, the marginal maximum posterior estimate formula of the variance with unbiased and efficient properties is derived.

Keywords

parameters / variance / statistical distribution / marginal distributions / maximum-posterior estimation

References

[1] YuZongchou, YuZhenglinThe principle of surveying adjustment, 1990, Wuhan, the publishing house of Wuhan Technical University of Surveying and Mapping in Chinese: 36-150

[2] CuiXizhang, YuZongchou, TaoBenzao, et al.The generaliced surveying adjustment, 1985, Beijing, the publishing house of surveying and mopping in chinese: 31-33

[3] ChengPing, ChenXiru, ChenGuijing, et al.The estimate of parameter, 1985, Shanghai, the publishing house of Shanghai Science and Technology in Chinese: 123-161

[4] WangSongguiThe theory and its application of linear model, 1987, Hefei, The publishing house of Anhui Education in Chinese: 53-63

[5] Edward M, Mikhail, F Ackermann. Observations and least squares. New York: [s. n.], 1976. 333–359

[6] HuangWeibinThe theory and its application of modern adjustment, 1992, Beijing, The publishing house of the PAL in Chinese: 74-84