摘要
针对非随机缺失机制下的数据缺失问题,在总体服从两参数Gamma分布和成功概率为指数分布的假设下,结合矩法和迭代计算方法进行了两参数Gamma分布模型的参数估计和分位数估计。此外,利用R语言对缺失比例分别为0.2和0.5的数据进行了模拟研究,并对模拟结果进行了比较分析,得到了较好的结果。
This paper focuses on the problem of data missing under non-random missing mechanism. Under the assumption that the population distribution follows two-parameter Gamma and the probability of success was exponential distribution,the parametric estimation and quantiles of two-parameter Gamma distribution model were obtained via moment method and iterative computation. Simulation studies were carried out using R language,when the missing ratio takes value 0. 2 and 0. 5 respectively. The simulation results were compared and analyzed,which showed better performance.
作者
梁远胜
关宏建
Liang Yuansheng;Guan Hongjian(Department of Mathematics, Yangjiang Polytechnic, Yangjiang Guangdong 529566;School of Mathematics and Statistics, Central China Normal University, Wuhan Hubei 430079)
出处
《湖北理工学院学报》
2018年第2期56-60,共5页
Journal of Hubei Polytechnic University
关键词
缺失数据
GAMMA分布
参数估计
矩估计
分位数
missing data
Gamma distribution
parametric estimation
moment estimator
quantile
