摘要
1非随机窗宽情形下的核估计设T_k表示某系统中第k个部件的寿命,整个系统的寿命为T=min{T_1,…,T_r},用R={1,2,…,r}表示风险集合。用下述ζ(T)表示系统的失效模式ζ(T)=,若T=T_j,j∈I,且T≠T_j,jI ,其它这里T=(T_1,T_2,…,T_r),I为R的一个子集。又设φ为R的非空子集的集合,φ_1={J∈φ,J∩I≠={J∈φ,J∩I=},则竞争风险问题即为估计2′-1个生存函数:S_1(l)
The kernel estimates of density functions is discussed under the competing risks case where T_1, T_2 …, T, may not be independent random variables. Asymptotic properties of two estimators, arising naturally as a result of considering two types of band widths, are investigated. In particular we show that the strong uniform consistency of the kernel estimates.
出处
《东南大学学报(自然科学版)》
EI
CAS
CSCD
1993年第1期113-117,共5页
Journal of Southeast University:Natural Science Edition
基金
东南大学青年科学基金
关键词
竞争风险
密度函数
核估计
non-parametric statistics / density function, competing risks, kernel estimate, strong consistency