Non-linear regression is an important tool in econometric and biometric applications. The goal of this master thesis is to apply the volume of the tube method based on the Weyl’s tube formula to construct simultaneous confidence bands for predicted regression curves in a non-linear regression setting with iid errors that are elliptically distributed. The outcome of the method has to be compared with confidence regions obtained by means of other approximation techniques (as, e.g., the bootstrap) in an extensive simulation study.

This thesis will be supervised jointly by the Institutes of Statistics and Stochastics.

Back to general view.