Sankhya: The Indian Journal of Statistics

1998, Volume 60, Series A, Pt. 3, pp. 446-458

DESIGNING OBSERVATION TIMES FOR INTERVAL CENSORED DATA

By

GIOVANNI PARMIGIANI, *Duke University, Durham*

*SUMMARY. *This paper studies the optimal choice of observation times for duration
data, presenting necessary conditions for the optimality of a sequence
of observation times under two alternative criteria, and characterizing
the form of the solution in special cases. A corollary of these results
is a simple expression for the binary partition of a continuous random
variable that is most informative for learning about an unknown parameter
in its distribution.

*AMS (1991) subject classification. *62F15.

*Key words and phrases. *Survival data, Kullback-Leibler information
criterion, Bayesian design of experiments, exponential distribution.