删失数据( censored data)由于某种原因被截断了的数据。如:灯泡在500小时的寿命试验中没有毁坏;无线电信号在受到强烈干扰而无法接收前已经持续了15分钟;吃了致癌物质的小白鼠到4月2日早晨8时尚无明显反应;机器从今天上午6时已经出了3次毛病等。按删失方式分,分为第一类删失、第二类删失和随机删失;按时间延伸方向分,分为左删失、右删失和区间删失。
但计算很是困难,原因就是删失数据后样本的似然函数的形式很复杂。
But it was difficult to calculate, because censored data in the form of samples of the likelihood function is very complex.
传统的乘积限估计法只能处理寿终数据和右删失数据,对左截断数据则无能为力。
The traditional product limit estimation method can only deal with complete data and right censored data, but can not deal with left truncated data.
基于K型区间删失数据,利用样本空间排序法给出参数优良的置信下限和计算置信下限的递推公式。
In the case of K type interval censored data, the lower confidence limit of parameter is studied based on the order relation established in the sample space.
应用推荐