CMStatistics 2016: Start Registration
View Submission - CMStatistics
B1524
Title: Size on sets of indistinguishable models Authors:  Anne Balter - Tilburg University (Netherlands) [presenting]
Antoon Pelsser - Maastricht University (Netherlands)
Abstract: Models can be wrong and recognising their limitations is important in financial and economic decision making under uncertainty. Finding the explicit specification of the uncertainty set has been difficult so far. We develop a method that provides a plausible set of models to use in robust decision making. The choice of the specific size of the uncertainty region is what we will focus on. We use the Neyman-Pearson Lemma to characterise a set of models that cannot be distinguished statistically from a baseline model. The set of indistinguishable models can explicitly be obtained for a given probability for the Type I and II error.