Network


Latest external collaboration on country level. Dive into details by clicking on the dots.

Hotspot


Dive into the research topics where Sebastian Döhler is active.

Publication


Featured researches published by Sebastian Döhler.


Statistical Inference for Stochastic Processes | 2003

Nonparametric estimation of regression functions in point process models

Sebastian Döhler; Ludger Rüschendorf

We prove that the empirical L2-risk minimizing estimator over some general type of sieve classes is universally, strongly consistent for the regression function in a class of point process models of Poissonian type (random sampling processes). The universal consistency result needs weak assumptions on the underlying distributions and regression functions. It applies in particular to neural net classes and to radial basis function nets. For the estimation of the intensity functions of a Poisson process a similar technique yields consistency of the sieved maximum likelihood estimator for some general sieve classes.


Statistics & Probability Letters | 2001

An approximation result for nets in functional estimation

Sebastian Döhler; Ludger Rüschendorf

In this paper a quantitative approximation result is obtained for a general class of function nets which is of interest in functional estimation. Specific applications are given to approximation by neural nets, radial basis function nets, and wavelet nets. For the proof we combine the empirical process based results of a paper of Yukich et al. (IEEE Trans. Inform. Theory 41 (4) (1995) 1021) with probabilistic based approximation results of Makovoz (J. Approx. Theory 85 (1996) 98) for the optimal approximation of functions by convex combination of n basis elements.


Statistical Methods in Medical Research | 2017

An extended sequential goodness-of-fit multiple testing method for discrete data.

Irene Castro-Conde; Sebastian Döhler; Jacobo de Uña-Álvarez

The sequential goodness-of-fit (SGoF) multiple testing method has recently been proposed as an alternative to the familywise error rate- and the false discovery rate-controlling procedures in high-dimensional problems. For discrete data, the SGoF method may be very conservative. In this paper, we introduce an alternative SGoF-type procedure that takes into account the discreteness of the test statistics. Like the original SGoF, our new method provides weak control of the false discovery rate/familywise error rate but attains false discovery rate levels closer to the desired nominal level, and thus it is more powerful. We study the performance of this method in a simulation study and illustrate its application to a real pharmacovigilance data set.


The Journal of Risk Model Validation | 2010

Validation of credit default probabilities via multiple testing procedures

Sebastian Döhler

We apply multiple testing procedures to the validation of estimated default probabilities in credit rating systems. The goal is to identify rating classes for which the probability of default is estimated inaccurately, while still maintaining a predefined level of committing type I errors as measured by the familywise error rate (FWER) and the false discovery rate (FDR). For FWER, we also consider procedures that take possible discreteness of the data resp. test statistics into account. The performance of these methods is illustrated in a simulation setting and for empirical default data.


Archive | 2009

Auswahl und Überprüfung von Modellen

Claudia Cottin; Sebastian Döhler

In den vorangegangenen Kapiteln wurden verschiedene Modelle zur Beschreibung von Risiken formuliert. Beispielsweise wird im Black-Scholes-Modell fur Optionspreise (s. Abschnitt 3.2.2) angenommen, dass der Kurs des Basiswerts einer geometrischen Brownschen Bewegung folgt. Modelle liefern jedoch nur eine Annaherung an die Realitat und stellen daher stets einen Kompromiss zwischen Einfachheit und Vollstandigkeit dar. Ziel ist also die Entwicklung eines Risikomodells, das die fur eine konkrete Fragestellung relevanten Risikoaspekte gut beschreibt. Damit stellt sich die Frage, wie ein solches Modell ausgewahlt bzw. unpassende Modelle ausgeschlossen werden konnen.


Archive | 2009

Mathematische Modellierung von Risiken

Claudia Cottin; Sebastian Döhler

Ein Modell fur ein einzelnes Risiko oder auch fur ein aus mehreren Einzelrisiken resultierendes Gesamtrisiko besteht im Kern aus Annahmen zur Wahrscheinlichkeitsverteilung; vgl. Kapitel 1. Fur ein ausgefeiltes privates oder unternehmerisches Risikomanagement ist zwar eine Gesamtbetrachtung aller wichtigen Risiken erstrebenswert. Allerdings sind dafur geeignete Modelle oft sehr komplex, auch wenn viele Vereinfachungen vorgenommen werden. Es empfiehlt sich also, zunachst mit der Analyse bzw. Modellierung von Einzelrisiken zu beginnen und ggf. in einem weiteren Schritt gleichartige Risiken zusammenzufassen. Was als ein einzelnes Risiko angesehen wird, hangt vom Kontext bzw. dem Detaillierungsgrad der Modellierung ab. Beispielsweise kann das, wenn es um Feuerschaden geht, ein einzelnes Gebaude sein oder der gesamte Gebaudebestand eines Unternehmens, oder aber das einzelne Risiko bezieht sich auf alle Arten moglicher Schaden (durch Feuer, Wasser, Sturm usw.), die an einem Gebaude entstehen konnen. Auch der betrachtete Zeithorizont spielt eine Rolle.


Statistics | 2003

A consistency result in general censoring models

Sebastian Döhler; Ludger Rüschendorf

In this paper we prove a consistency result for sieved maximum likelihood estimators of the density in general random censoring models with covariates. The proof is based on the method of functional estimation. The estimation error is decomposed in a deterministic approximation error and the stochastic estimation error. The main part of the proof is to establish a uniform law of large numbers for the conditional log-likelihood functional, by using results and techniques from empirical process theory.


Archive | 2003

On Adaptive Estimation by Neural Net Type Estimators

Sebastian Döhler; Ludger Rüschendorf

We investigate the quality of neural net based estimators in the model example of estimating the conditional log-hazard function in censoring models. The quality of estimators is measured in terms of convergence rates. Our bounds for the convergence rate show that for some activation functions like the threshold function and for piecewiese polynomial functions optimal convergence rates are attained by the corresponding neural net estimators. For the standard sigmoid function we however obtain only bounds which indicate suboptimal behavior. A complexity regularized version of the neural net type estimators is also considered and is shown to be approximatively adaptive in smoothness classes.


Econometrics and Statistics | 2018

A discrete modification of the Benjamini–Yekutieli procedure

Sebastian Döhler


Statistics & Probability Letters | 2014

A sufficient criterion for control of some generalized error rates in multiple testing

Sebastian Döhler

Collaboration


Dive into the Sebastian Döhler's collaboration.

Top Co-Authors

Avatar
Top Co-Authors

Avatar
Top Co-Authors

Avatar
Top Co-Authors

Avatar
Top Co-Authors

Avatar
Researchain Logo
Decentralizing Knowledge