Construction of Approximate Confidence Intervals for Probability-of-Kill via the Bootstrap

摘要

The bootstrap, a computer-intensive procedure for data analysis, was applied to an estimation problem to enable a statement to be made about the variability inherent in a probability-of-kill estimate P sub k. The bootstrap was applied to a stratified sample rather than a simple random sample and its performance evaluated in this framework, first in an abstract situation where the parameter P sub k was known and then in three situations where the parameter P sub k was unknown, but an estimate provide by current vulnerability analysis procedures was available. The investigation was carried out for several grid sizes, or alternatively, for several levels of detail, to study the effect of grid size on the estimation of P sub k and the reliability of the confidence intervals constructed. (Author)