摘要
对于部分线性模型y_i=βx_i+g(t_i)+e_i,1≤i≤n,这里(x_i,t_i)是固定设计点,g是未知函数,e_i是负相协(NA)随机误差,给出了回归系数的经验似然比统计量,并讨论了似然比统计量的极限分布,可构造参数的经验似然置信区间.
This paper is concerned with the typical partial linear model yi=βxi+g(ti)+ei,1≤i≤n, where (xi, ti) are fixed design points, g is an unknown function, and ei's are negatively associated (NA) random errors. An empirical log-likelihood ratio for the regression coefficient is proposed, the results show that the statistic is asymptotically chisquared distributed and that the confidence intervals can be constructed accordingly.
出处
《系统科学与数学》
CSCD
北大核心
2009年第4期490-501,共12页
Journal of Systems Science and Mathematical Sciences
基金
国家自然科学基金(10571073,J10630104)
高等学校博士学科点专项基金(20070183023)
吉林大学“985”工程
长春税务学院科学研究(2008016)项目资助.
关键词
部分线性模型
负相协随机误差
经验对数似然
置信区间.
Partial linear model, negatively associated random errors, empirical loglikelihood, confidence intervals.