NPML estimation of the mixing distribution in general statistical models with unobserved random effects

1995 ◽  
pp. 1-9
Author(s):  
Murray Aitkin
Keyword(s):  
Random Effects ◽  
Statistical Models ◽  
Mixing Distribution ◽  
General Statistical

Prior Information in Spatial Analysis

1978 ◽  
Vol 10 (1) ◽  
pp. 51-70 ◽  
Author(s):  
J Odland
Keyword(s):  
Spatial Data ◽  
Statistical Models ◽  
Prior Information ◽  
Population Distribution ◽  
Analytical Models ◽  
Distribution Model ◽  
Density Functions ◽  
Growth Processes ◽  
Statistical Procedures ◽  
General Statistical

Formal statistical procedures which incorporate prior, or nonsample, information can be used to enlarge the theoretical and empirical content of analytical models, and can be widely applied to adapt general statistical procedures for special investigative situations which arise in the analysis of spatial data. A class of methods which operate by modifying the parameter space of general statistical models in accordance with prior information is presented in this paper. The prior information can be obtained from theoretical or empirical sources, and the models can be given Bayesian as well as non-Bayesian interpretations. The utility of the methods is demonstrated by applying them to estimate parameters for urban-population-density functions. The resulting estimates incorporate nonsample information on total populations as well as sample information on subarea densities. The logic of the population-distribution model is also extended to take account of possible growth processes by incorporating prior information derived from earlier distributions and by incorporating the characteristics of possible growth processes.

scholarly journals Higher-order asymptotic theory of shrinkage estimation for general statistical models

2018 ◽  
Vol 166 ◽  
pp. 198-211
Author(s):  
Hiroshi Shiraishi ◽  
Masanobu Taniguchi ◽  
Takashi Yamashita
Keyword(s):  
Statistical Models ◽  
Asymptotic Theory ◽  
Higher Order ◽  
Shrinkage Estimation ◽  
General Statistical

PSVIII-4 Genetic evaluation of functional heifer longevity in north American angus cattle

2021 ◽  
Vol 99 (Supplement_3) ◽  
pp. 240-241
Author(s):  
Hinayah R Oliveira ◽  
Stephen P Miller ◽  
Luiz F Brito ◽  
Flavio S Schenkel
Keyword(s):  
Random Effects ◽  
Maternal Effects ◽  
Statistical Models ◽  
Evaluation System ◽  
Genetic Model ◽  
Genetic Evaluation ◽  
Genetic Effects ◽  
Reduced Model ◽  
Complete Model ◽  
Angus Cattle

Abstract The goals of this study were to develop a genetic evaluation system for a novel trait called functional heifer longevity (FHL), and determine if this novel trait is heritable. The FHL trait was defined as binary, in which the heifers received the code 1 if they had calved by the end of their third year (n = 377,938), or 0 if they were culled/sold during this period (n = 368,308). Analysis were performed using linear animal models and Bayesian inference. The significant systematic effects included in the statistical models are born by embryo transfer, year-season of birth, and age at calving (in months). Three models, differing according to their random effects (i.e., reduced model, which included only herd-year-season and additive genetic random effects; maternal genetic model, which added maternal genetic effects; and complete model, which further added maternal permanent environmental effects), were compared based on the deviance information criterion (DIC) and the estimates of genetic parameters. The reduced model was preferred according to the DIC values. However, high maternal heritabilities were estimated using the maternal genetic (0.51) and complete (0.36) models, indicating that maternal effects can impact the selection of heifers for breeding. Similar additive genetic heritabilities were estimated among the three models (0.24, 0.27, and 0.25 using the reduced, maternal genetic, and complete models, respectively), and no significant re-ranking of selection candidates were observed based on their additive genetic breeding values. Total heritabilities and correlations estimated between additive genetic and maternal genetic effects were 0.37 and -0.28 for the maternal genetic, and 0.31 and -0.27 for the complete model, respectively. This study shows that FHL is heritable, and that including maternal effects in the statistical models might be important. These results contribute to a larger project studying the genetics of female longevity in Angus cattle.

Random Effects Won’t Solve the Problem of Generalizability

2021 ◽  
Author(s):  
Adam Bear ◽  
Jonathan Scott Phillips
Keyword(s):  
Random Effects ◽  
Statistical Models ◽  
Sources Of Variation ◽  
Study Designs

Yarkoni argues that researchers making broad inferences often use impoverished statistical models that fail to include important sources of variation as random effects. We argue, however, that for many common study designs random effects are inappropriate and insufficient to draw general inferences, as the source of variation is not random, but systematic.

Asymptotic Expansions for General Statistical Models.

Economica ◽  
1987 ◽  
Vol 54 (214) ◽  
pp. 268
Author(s):  
P. M. Robinson ◽  
J. Pfanzagl ◽  
W. Wefelmeyer
Keyword(s):  
Asymptotic Expansions ◽  
Statistical Models ◽  
General Statistical

scholarly journals Designing Optimal Bonus-Malus Systems from Different Types of Claims

1998 ◽  
Vol 28 (2) ◽  
pp. 205-220 ◽  
Author(s):  
Jean Pinquet
Keyword(s):  
Random Effects ◽  
Poisson Model ◽  
Expected Value ◽  
Third Party ◽  
Mixing Distribution ◽  
Empirical Results ◽  
Different Types ◽  
Consistent Estimators

AbstractThis paper provides bonus-malus systems which rest on different types of claims. Consistent estimators are given for some moments of the mixing distribution of a multi equation Poisson model with random effects. Bonus-malus coefficients are then obtained with the expected value principle, and from linear credibility predictors. Empirical results are presented for two types of claims, namely claims at fault and not at fault with respect to a third party.

Sign in / Sign up

Export Citation Format

Share Document