Statistics Seminars: Optimal designs for threshold detection using limiting dilution assays
31 January 2011 15:15 in CM221
As with non-linear models generally, the optimal design for a limiting dilution assay depends on the unknown value of the parameter of interest. If the available information about the parameter can be encapsulated in a prior distribution then an optimal Bayesian design can be sought. If the aim is to estimate the parameter then various approximations to the full Bayesian criterion, analogous to the usual alphabetic optimality criteria, are often used. However, if the assay has a different purpose then more specific utility functions can provide improved designs. This is illustrated by applications of limiting dilution assays in which the aim of the study is to determine if the parameter is above or below a specified threshold.
Contact email@example.com for more information