带协变量的负二项回归模型中离散参数估计的研究

摘要

针对带协变量的负二项回归模型中离散参数估计问题,推广了极大似然估计和Bootstrap极大似然估计方法,并在绝对偏差的意义下,通过模拟研究和实际数据分析研究了估计的优良性.研究结果表明协变量和样本量均对离散参数估计有影响.%This paper extends the maximum likelihood and bootstrap maximum likelihood methods for dispersion parameter estimation in the negative binomial regression model with covariates.By a Monte Carlo simulation study and a real data analysis, the superiority of the proposed estimators is compared.Furthermore,the influence of covariates and sample size on the estimators of the dispersion parameter is found.