逆高斯分布是分析非对称右偏数据的重要统计工具,该文研究提出混合逆高斯分布下广义线性联合均值与散度模型,利用EM算法和Fisher-Scoring方法研究了模型参数极大似然估计.最后,通过随机模拟和实例研究结果表明,提出的模型和方法是有用和有效的.
The inverse Gaussian distribution is one of the most important statistical data analysis tools for positive skew data. A generalized linear joint mean and dispersion model based on mixture inverse Gaussian distribution is presented.The maximum likehood extimation of unknown parameters is researched using EM algorithm and Fisher-Scoring method.Simulation and practice show that these models and meth- ods are useful and effective.