摘要
首先,利用核光滑方法研究左截断右删失数据下总体分位数差的估计,得到了左截断右删失数据下分位数差的光滑估计及估计量的大样本性质.其次,在均方误差意义下,证明了光滑分位数差估计比左截断右删失数据下乘积限分位函数的差有更高的估计效率.最后数值模拟分析高斯核函数下选择不同窗宽对改善乘积限分位数差估计效率的影响.
Firstly,by using the kernel smooth method,we studied the estimation of the total quantile difference with left-truncated and right-censored(LTRC)data,and obtained the large sample properties of the smoothed estimators of quantile difference with LTRC data.Secondly,in the sense of mean square error,we proved the estimation efficien cy of the smooth quantile difference was higher than that of the pro duct limit quantile difference with LTRC data.Finally,we analyzed the effect o f various bandwidths with Gaussian kernel function on the improved estimation ef ficiency of product limit quantile difference through numerical simulations.
作者
荀立
崔世崇
朵兰
XUN Li;CUI Shichong;DUO Lan(School of Mathematics and Statistics,Changchun University of Technology,Chang chun 130012,China)
出处
《吉林大学学报（理学版）》
CAS
CSCD
北大核心
2018年第5期1105-1112,共8页
Journal of Jilin University:Science Edition
基金
国家自然科学基金(批准号:11701043)
关键词
左截断右删失数据
光滑分位函数
乘积限分位函数
分位数差
left-truncated and right-censored(LTRC)data
smooth quan tile function
product limit quantile function
quantile difference
