广义Logistic回归模型Bayes分析及其在林木存活率预报中的应用

Bayes analysis for generalized Logistic regression model and its application to forestry survival rate

通过因子分析的方法来解释观测变量的相关性,这种方法是通过引入潜在变量来直接刻画相关性。基于Bayes统计原理、方法用来解决模型的参数估计问题和统计推断,采用Markov Chains Monte Carlo(MCMC)进行统计计算。随机模拟的结果表明所提出的林木存活率预测方法是有效的。最后利用该方法对山西沁水县沁水林场的林木存活率与林分的鼠兔数关系进行了分析。

Factor analysis, which is characterized by the latent variables, is a popular method to interpret the correlation among the observed variables. In this paper, latent constructs are introduced to describe the relationship of the categorical variables. Within the Bayesian framework, parameters estimations and statistical inferences are carried out via a popular technique, i.e., Markov Chains Monte Carlo （MCMC）. A simulation study is conducted to assess the proposed method. A pika data set is used to illustrate the real application.