Department of Dairy Science, University of Wisconsin, Madison, 53706, USA.
J Anim Breed Genet. 2010 Feb;127(1):3-15. doi: 10.1111/j.1439-0388.2009.00835.x.
Structural equation models provide a general statistical modelling technique for estimating and testing relationships among variables. Such relationships are often not revealed by standard linear models, but are of importance for understanding mechanisms underlying e.g., production-related diseases, such as mastitis. This paper gives a review of Bayesian structural equation models concerning methodology and identifiability, focused on animal breeding and genetics modelling. Applications of this type of methods in animal breeding are also reviewed critically, with discussion on advantages and disadvantages of these approaches.
结构方程模型为估计和检验变量之间的关系提供了一种通用的统计建模技术。这种关系通常不是标准线性模型所揭示的，但对于理解例如与生产相关的疾病（如乳腺炎）的机制很重要。本文回顾了贝叶斯结构方程模型在方法和可识别性方面的内容，重点是动物育种和遗传学建模。还批判性地回顾了这种方法在动物育种中的应用，并讨论了这些方法的优缺点。
