• Safety and Health at Work
• /
• 제12권3호
• /
• pp.296-303
• /
• 2021

Background: Rapid population aging in developed countries has resulted in the working-age population increasingly being tasked with the provision of informal care. Methods: An educational intervention was delivered to 21 carer-employees employed at a Canadian University. Work role function, job security, schedule control, work-family conflict, familywork conflict, and supervisor and coworker support were measured as part of an aggregated workplace experience score. This score was used to measure changes pre/post intervention and at a follow-up period approximately 12 months post intervention. Three random intercept models were created via linear mixed modeling to illustrate changes in participants' workplace experience across time. Results: All three models reported statistically significant random and fixed effects intercepts, with a positive coefficient of change. Conclusion: This suggests that the intervention demonstrated an improvement of the workplace experience score for participants over time, with the association particularly strong immediately after intervention.

• 응용통계연구
• /
• 제23권6호
• /
• pp.1169-1178
• /
• 2010

본 논문은 불균형 자료에서 선형혼합모형에 적용되는 Akaike Information Criterion(AIC)의 효율에 대한 연구이다. Vaida와 Balanchard (2005)에 의해 제안된 cAIC(conditional AIC)는 mAIC(marginal AIC)가 임의효과의 예측에 대한 불확실성을 모형선택에서 반영하지 못하는 단점을 극복할 수 있는 방법이다. cAIC에 대한 이론적인 성질과 확장은 Liang 등 (2008)과 Greven과 Kneib (2010)에 의하여 연구되었다. cAIC의 형태는 자료의 구조에 영향을 받지는 않지만 선형혼합모형에서 모수의 추정 효율은 자료의 불균형의 정도에 따라 많은 영향을 받는 것이 알려져 있다. 기존의 연구에서 실시한 모든 모의실험이 자료가 균형인 경우에만 실행되어 자료의 불균형이 AIC에 근거한 혼합모형 선택 방법의 효율에 어떤 영향을 미치는지 알려져 있지 않다. 본 논문은 자료의 불균형이 모형선택 방법의 효율에 미치는 영향을 모의실험을 통하여 알아보았다. 자료의 불균형이 심해짐에 따라 AIC에 근거한 모형선택방법은 복잡한 모형을 선택하는 경향이 낮아짐을 보였다.

• Asian-Australasian Journal of Animal Sciences
• /
• 제15권7호
• /
• pp.925-931
• /
• 2002

Gibbs sampling algorithms were implemented to the multi-trait threshold animal models with any combinations of multiple binary, ordered categorical, and linear traits and investigate the amount of bias on these models with two kinds of parameterization and algorithms for generating underlying liabilities. Statistical models which included additive genetic and residual effects as random and contemporary group effects as fixed were considered on the models using simulated data. The fully conditional posterior means of heritabilities and genetic (residual) correlations were calculated from 1,000 samples retained every 10th samples after 15,000 samples discarded as "burn-in" period. Under the models considered, several combinations of three traits with binary, multiple ordered categories, and continuous were analyzed. Five replicates were carried out. Estimates for heritabilities and genetic (residual) correlations as the posterior means were unbiased when underlying liabilities for a categorical trait were generated given by underlying liabilities of the other traits and threshold estimates were rescaled. Otherwise, when parameterizing threshold of zero and residual variance of one for binary traits, heritability estimates were inflated 7-10% upward. Genetic correlation estimates were biased upward if positively correlated and downward if negatively correlated when underling liabilities were generated without accounting for correlated traits on prior information. Residual correlation estimates were, consequently, much biased downward if positively correlated and upward if negatively correlated in that case. The more categorical trait had categories, the better mixing rate was shown.

• Journal of the Korean Statistical Society
• /
• 제31권2호
• /
• pp.177-198
• /
• 2002

The estimation of variance components or variance ratios in linear model is an important issue in plant or animal breeding fields, and various estimation methods have been devised to estimate variance components or variance ratios. However, many traits of economic importance in those fields are observed as dichotomous or polychotomous outcomes. The usual estimation methods might not be appropriate for these cases. Recently threshold linear model is considered as an important tool to analyze discrete traits specially in animal breeding field. In this note, we consider a hierarchical Bayesian method for the threshold animal model. Gibbs sampler for making full Bayesian inferences about random effects as well as fixed effects is described to analyze jointly discrete traits and continuous traits. Numerical example of the model with two discrete ordered categorical traits, calving ease of calves from born by heifer and calving ease of calf from born by cow, and one normally distributed trait, birth weight, is provided.

• Communications for Statistical Applications and Methods
• /
• 제10권1호
• /
• pp.11-18
• /
• 2003

Quite often normality assumptions are not satisfied in practical applications. In this paper, an estimated generalized least squares(EGLS) analysis are considered in two way mixed linear models with arbitrary types of distributions for random effects. We investigate the power performance of EGLS analysis based on Henderson's method III, ML, REML and MINQUE(1). The power performances depend on the imbalance of design, on the actual values of ratio of variance components, and on the skewness and kurtosis parameters of the underlying distributions slightly. Results of our limited simulation study suggest that the EGLS F-statistics using four estimators and arbitrary distributions produce similar type I error rates and power performance.

• Journal of the Korean Data and Information Science Society
• /
• 제27권5호
• /
• pp.1389-1397
• /
• 2016

Semi-parametric frailty model with nonparametric baseline hazards has been widely used for the analyses of clustered survival-time data. The frailty models can be fitted via an auxiliary Poisson hierarchical generalized linear model (HGLM). For the inferences of the frailty model marginal likelihood, which gives MLE, is often used. The marginal likelihood is usually obtained by integrating out random effects, but it often requires an intractable integration. In this paper, we propose to obtain the MLE via Laplace approximation using a Poisson HGLM approach for semi-parametric frailty model. The proposed HGLM approach uses hierarchical-likelihood (h-likelihood), which avoids integration itself. The proposed method is illustrated using a numerical study.

• East Asian mathematical journal
• /
• 제5권1호
• /
• pp.95-110
• /
• 1989

In a general variance component model, nonnegative quadratic estimators of the components of variance are considered which are invariant with respect to mean value translaion and have minimum bias (analogously to estimation theory of mean value parameters). Here the minimum is taken over an appropriate cone of positive semidefinite matrices, after having made a reduction by invariance. Among these estimators, which always exist the one of minimum norm is characterized. This characterization is achieved by systems of necessary and sufficient condition, and by a cone restricted pseudoinverse. In models where the decomposing covariance matrices span a commutative quadratic subspace, a representation of the considered estimator is derived that requires merely to solve an ordinary convex quadratic optimization problem. As an example, we present the two way nested classification random model. An unbiased estimator is derived for the mean squared error of any unbiased or biased estimator that is expressible as a linear combination of independent sums of squares. Further, it is shown that, for the classical balanced variance component models, this estimator is the best invariant unbiased estimator, for the variance of the ANOVA estimator and for the mean squared error of the nonnegative minimum biased estimator. As an example, the balanced two way nested classification model with ramdom effects if considered.

• Animal Bioscience
• /
• 제34권2호
• /
• pp.163-171
• /
• 2021

Objective: Considering the importance of dairy farming and the negative effects of heat stress, more tolerant genotypes need to be identified. The objective of this study was to investigate the effect of heat stress via temperature-humidity index (THI) and diurnal temperature variation (DTV) in the genetic evaluations for daily milk yield of Holstein dairy cattle, using random regression models. Methods: The data comprised 94,549 test-day records of 11,294 first parity Holstein cows from Brazil, collected from 1997 to 2013, and bioclimatic data (THI and DTV) from 18 weather stations. Least square linear regression models were used to determine the THI and DTV thresholds for milk yield losses caused by heat stress. In addition to the standard model (SM, without bioclimatic variables), THI and DTV were combined in various ways and tested for different days, totaling 41 models. Results: The THI and DTV thresholds for milk yield losses was THI = 74 (-0.106 kg/d/THI) and DTV = 13 (-0.045 kg/d/DTV). The model that included THI and DTV as fixed effects, considering the two-day average, presented better fit (-2logL, Akaike information criterion, and Bayesian information criterion). The estimated breeding values (EBVs) and the reliabilities of the EBVs improved when using this model. Conclusion: Sires are re-ranking when heat stress indicators are included in the model. Genetic evaluation using the mean of two days of THI and DTV as fixed effect, improved EBVs and EBVs reliability.

• Asian-Australasian Journal of Animal Sciences
• /
• 제19권4호
• /
• pp.468-474
• /
• 2006

Real time ultrasound data was generated on 10,596 live Hanwoo cows to study genetic variation on ultrasonic beef quality traits and to assess the best model to estimate genetic parameters on these traits. Pedigree stacking and data validation was done using the SAS statistical software and the genetic parameter estimates were obtained by EM-REML algorithm. Out of the five different multi-trait mixed animal models constructed, the optimal model included fixed effects of herd, year-season-appraisal, body condition score, linear and quadratic covariates for chest girth, the linear covariate effect of age and the random animal and residual effect of the five models studied. The heritability of longissimus muscle area (LMA), $12^{th}$ rib measurement of back fat thickness (BF) and marbling score (MS) was 0.11, 0.17 and 0.15, respectively. Genetic correlation of LMA vs. BF, LMA vs. MS and BF vs. MS was -0.15, 0.06 and 0.61, respectively. The results showed presence of genetic variation in these ultrasonic beef quality traits in Hanwoo cows and suggest that the selection of Hanwoo cows may be possible by performing ultrasonic scans on live animals, which will ultimately be helpful in reducing the generation interval and the cost of selection procedure.

• Communications for Statistical Applications and Methods
• /
• 제29권2호
• /
• pp.177-191
• /
• 2022

This paper addresses the use of machine learning methods for causal estimation of treatment effects from observational data. Even though conducting randomized experimental trials is a gold standard to reveal potential causal relationships, observational study is another rich source for investigation of exposure effects, for example, in the research of comparative effectiveness and safety of treatments, where the causal effect can be identified if covariates contain all confounding variables. In this context, statistical regression models for the expected outcome and the probability of treatment are often imposed, which can be combined in a clever way to yield more efficient and robust causal estimators. Recently, targeted maximum likelihood estimation and causal random forest is proposed and extensively studied for the use of data-adaptive regression in estimation of causal inference parameters. Machine learning methods are a natural choice in these settings to improve the quality of the final estimate of the treatment effect. We explore how we can adapt the design and training of several machine learning algorithms for causal inference and study their finite-sample performance through simulation experiments under various scenarios. Application to the percutaneous coronary intervention (PCI) data shows that these adaptations can improve simple linear regression-based methods.