CMStatistics 2016: Start Registration
View Submission - CMStatistics
B0621
Title: Computing $c$-optimal designs for non-independent observations Authors:  Juan M Rodriguez-Diaz - University of Salamanca, CIF Q3718001E (Spain) [presenting]
Abstract: $c$-optimality is one of the most employed optimality criteria. For a given model it looks for the design that minimizes the variance of the linear combinationof the parameters' estimators given by vector $c$. $c$-optimal designs are needed when dealing with standardized criteria, and are frequently used to check how good a design is for the estimation of each of the parameters of the model. The Elfving's procedure for independent observations gives the idea of the procedure for correlated observations. Some results can be obtained for simple covariance structures; however when moving from these situations the computations become harder. An analysis of covariance structures suitable for the application of the procedure is performed, and its behavior will be illustrated with convenient examples.