摘要
考虑了响应变量随机缺失情形下的线性EV模型,构造了响应变量均值的两种经验对数似然比统计量,即估计的经验对数似然比统计量和调整的经验对数似然比统计量,证明了所构造的经验似然比统计量渐近于χ2分布,所得结果可以用来构造置信区间.
A linear EV model with missing data is investigated.Two empirical log-likelihood ratio statistics for the mean of response variable are proposed.It is proved that the proposed statistics are asymptotically chi-square under some mild conditions,and hence can be used to construct the confidence intervals.
出处
《广西师范学院学报(自然科学版)》
2010年第3期24-26,共3页
Journal of Guangxi Teachers Education University(Natural Science Edition)
基金
广西自然科学基金项目(0832108
2010GXNSFB013048)
关键词
线性EV模型
缺失数据
经验似然
置信区间
linear EV model
missing data
empirical likelihood
confidence interval