November 15th 2013
Faculty of Economic and Business Sciences | Salón de Grados
Controlling generalised error rates
2013/11/15 – 12:00 h | Sebastian Doehler, Darmstadt University of Applied Sciences
Abstract
Generalised error rates such as the k-familywise error rate and the false discovery proportion have been introduced in order to obtain more powerful multiple testing procedures. In this talk, I will present a simple sufficient criterion for controlling generalized error rates for arbitrarily dependent p-values. This criterion is formulated in terms of matrices associated with these error rates and thus it is possible to view the resulting mutliple testing procedures as solutions of sets of certain linear inequalities. This property can in some cases be used to improve the power of existing procedures by finding optimal solutions to a related linear programming problem. A simulation study and analyses of some empirical data illustrate the improved power of these modified procedures.