Dienstag, 28.04.2015, 12-13 Uhr - Raum: W9-109

Dr. Christian Schellhase
Universität Bielefeld

Semi-parametric Estimation of Income Mobility with D-vines using Bivariate Penalized Splines

The talk presents a semiparametric copula estimation approach for D-vines to evaluate the distribution of income mobility in Germany. We compare earning differentials on a micro level between four years in in the 1980's, 1990's and 2000's. We thereby separate with respect to the educational attainment of the individuals. We model income mobility by decomposing the multivariate distribution of invidual incomes into its univariate margins with bivariate copulas. The latter is carried out by fitting D-vines in a flexible semi-parametric way. Whereas D-vines are commonly used to model serial dependence in multivariate time series. We employ penalized Bernstein polynomials as spline basis. The penalty induce smoothness of the fit while the high dimensional spline basis guarantees flexibility. We use the Sample of Integrated Labour Market Biographies (SIAB) as official empirical data from the German Federal Employment Agency. As result, we detect very low income mobility for the low educated workforce throughout the decades. For the high educated individuals, we detect significantly higher income mobility. Due to the assumptions of D-vines, the latter is visualized by the corresponding conditional probabilities of income changes throughout the time periods. 

 

Dienstag, 12.05.2015, 12-13 Uhr - Raum: W9-109

Prof. Dr. Dietmar Bauer
Universität Bielefeld

Prognose von Routenfahrzeiten auf der Basis von Floating Taxi Daten

In diesem Vortrag werden verschiedene Ergebnisse zur Modellierung von Fahrzeiten in einem Verkehrsnetz gesammelt und aufbereitet. Ausgehend von einer Beschreibung der zugrunde liegenden Datenstruktur wird zunächst eine deskriptive Analyse einiger Segmentfahrzeiten-Profile durchgeführt. Anschließend werden einige Fakten zur Modellierung der erwarteten Reisezeit präsentiert. Den Abschluss des Vortrags bilden Resultate zur Unsicherheit dieser Prognosen sowie deren Verknüpfung für die Prognose von Routenreisezeiten besprochen.

 

Dienstag, 26.05.2015, 12-13 Uhr - Raum: W9-109

Prof. Dr. Göran Kauermann
Institut für Statistik der Ludwig-Maximilians-Universität München

Classification in Underwater Sonar Videos

Dual-frequency Identification Sonar (so called DIDSON) delivers video-like underwater images which allow to investigate fish behaviour even in cloudy and muddy water. Generally, images are recorded in a resolution of up to 10 pictures per second, so that practically one obtains a video of underwater movements. These videos allow ecologists to observe, count or investigate fish behaviour. In this talk we focus on automatic classification of fish based on such sonar videos. We show how after appropriate data preprocessing of the videos we can count and classify fish into different species based on their shape and movement. The developed procedures are working in real time, that is data processing and classification of video sequences is faster than the length of the video sequences itself.

 

Dienstag, 09.06.2015, 12-13 Uhr - Raum: W9-109

Philipp Külpmann
Institut für Mathematische Wirtschaftsforschung, Universität Bielefeld

Probabilistic Transitivity in Sports

We seek to find the statistical model that most accurately describes empirically observed results in sports. The idea of a transitive relation concerning the team strengths is implemented by imposing a set of constraints on the outcome probabilities. We theoretically investigate the resulting optimization problem and draw comparisons to similar problems from the existing literature including the linear ordering problem and the isotonic regression problem. Our optimization problem turns out to be very complicated to solve. We propose a branch and bound algorithm for an exact solution and for larger sets of teams a heuristic method for quickly finding a „good“ solution. Finally we apply the described methods to panel data from soccer, American football and tennis and also use our framework to compare the performance of empirically applied ranking schemes. Paper: http://ssrn.com/abstract=2490457

 

Dienstag, 23.06.2015, 12-13 Uhr - Raum: W9-109

Prof. Dr. Ostap Okhrin
Technische Universität Dresden

Efficient Iterative Maximum Likelihood Estimation

We propose an algorithm to efficiently estimate models with complex log-likelihood functions. Given a consistent but inefficient estimate of the parameter, the procedure yields a computationally tractable, consistent and asymptotically efficient estimate. We derive the estimator's asymptotic distribution with the asymptotic covariance depending on the number of iteration steps. We provide a rule of thumb to approximate the number of iterations until the estimator converges to the ordinary maximum likelihood estimator and discuss possibilities to accelerate the speed of the algorithm, e.g., by simplifying the model complexity. Small sample properties of the estimator are illustrated in a comprehensive simulation study. In an empirical application, we use the proposed method to estimate the volatility connectedness between companies by extending the approach by Diebold and Yilmaz (2014) to a non-Gaussian setting.

 

Dienstag, 07.07.2015, 12-13 Uhr - Raum: W9-109

Aktuelle Forschungsbereiche im ZeSt