Dear Allstaters,
Let Xi (Yi) represent failure (censored) times.
In "random censoring model", we observe the pairs:
(T1,Z1), (T2,Z2), ..., (Tn,Zn)
where Ti = min(Xi,Yi) and Zi=0(1) if Xi>Yi (otherwise)
The following are some special cases of the above model:
(1) If Y=infinty a.s., we obtain complete data
(2) If Yi=ri a.s. where ri are fixed positive constants,
we obtain progressively type I censored data at the points ri
(3) If ri=r for al i, we obtain "simple" type I censored data
My question now is:
Can other right censoring schemes (such as "simple" type II censoring and
progressively type II censoring) be considered also as special cases of the
random censoring model? If not, why?
Thanks in advance.
M.E. Ghitany
17/11/99
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|