摘要
计数数据分析已在生物和医学研究领域得到发展.特别地,零膨胀计数分布已经被用来建模在试验中经常出现的过多零的观测.分析此类数据最常见的计数分布是泊松分布和负二项分布.然而,泊松分布只能处理等分布数据,负二项分布只能处理过度分散.零膨胀广义Poisson回归模型则可以用来处理0过多或过度分散的计数数据.对于具有这样特征的纵向计数数据采用广义估计方程方法估计模型参数,同时也可以解释来自同一个体观测值之间的相关性.通过引入一个隐变量,将分别构建模型的两组协变量的估计方程连接起来,从而解决了零膨胀部分的参数估计求解问题.并用此方法分析了Iowa氟化物研究中的数据,研究了龋齿的数量影响因素.
Count data analysis techniques have been developed in biological and medical research areas.In particular,zero-inflated versions of parametric count distributions have been used to model excessive zeros that are often present in these assays.The most common count distributions for analyzing such data are Poisson and negative binomial.The zero-inflation generalized Poisson regression model is used to explain the problem of excessive or excessive dispersion of count data.For longitudinal counting data with such characteristics,the generalized estimation equation method is used to estimate the model parameters,and it can also explain the correlation between observations from the same individual.By introducing a potential variables,a model of two groups of covariate estimating equations connected,so as to solve the zero expansion part of the solution of the parameter estimation.And using this method to analyze the data in the Iowa fluoride studies,to explore the factors affecting the number of dental caries.
作者
殷明娥
于洋
YIN Ming’e;YU Yang(School of Mathematics, Liaoning Normal University, Dalian 116029, China)
出处
《辽宁师范大学学报（自然科学版）》
CAS
2020年第4期447-454,共8页
Journal of Liaoning Normal University:Natural Science Edition
基金
辽宁省科技厅引导项目(20180550196)。
关键词
广义估计方程
零膨胀广义Poisson回归模型
纵向数据
generalized estimating equation
zero-inflated generalized Poisson regression model
longitudinal data
