Dr. Alexander Kreiß

Contact

Email: alexander.kreiss@math.uni-leipzig.de
Phone: +49 (0) 341-9732-328
Office: A437, Augusteum
Office Hours: By appointment (just send me an email)

You can also find me on Github and Research Gate.

CV and Research Interests

In general my research interests are mostly centred around the analysis of Statistical Models for Network Data, particularly, Semi- and Non-parametric Methods, Dependence, Relational Event Models using counting processes, and Bootstrap methods. In addition, I'm interested in Causal Inference, High-Dimensional Statistics, Quantile Regression, Measurement Error Problems, Hawkes Processes and Survival Analysis. It is particularly interesting if some of these areas overlap. I'm scientific member of the International Max Planck Research school (IMPRS) of the Max Planck Institut for Mathematics in the Sciences, I'm a co-organizer of the Seminar for Statistical Learning and Probability/Dynamical Systems and I'm affiliated with the HKMetrics Network.

Before coming to Leipzig, I was a PostDoc at the University of Mannheim in the group of Prof. Christoph Rothe (2019-2020), at KU Leuven in the group of Prof. Ingrid Van Keilegom (2020-2021) and at LSE in the group of Prof. Qiwei Yao (2020-2021).
I've studied Mathematics with minor subject Electronics at TU Darmstadt (Bachelor 2012, Master 2015) and at the University of Bristol (Master 2014). From October 2015 to April 2019 I was a PhD student at Heidelberg University under the supervision of Prof. Enno Mammen (Heidelberg) and Prof. Wolfgang Polonik (UC Davis). In my thesis Local Maximum Likelihood Estimation of Time Dependent Parameters in Dynamic Interaction Networks I've studied non-parametric estimation of the parameter function in a model for interactions among actors in a network. The results of my thesis have been published in two papers (Electronic Journal of Statistics and Bernoulli, see below). I was member of the RTG 1953 from 2015 to 2018.

Current Projects and Pre-Prints

• Sparse Network Estimation Using Hawkes Processes (preliminary title), with Enno Mammen (Heidelberg University) and Wolfgang Polonik (UC Davis).
• Estimating Average Treatment Effects in Regression Discontinuity Designs with Covariates under Minimal Assumptions, with Patrick Kramer (under review)
• Efficient Quantile Regression under Censoring Using Laguerre Polynomials, with Ingrid Van Keilegom (Revision) - R-Code

Refereed Journals

• Testing For a Parametric Baseline-Intensity in Dynamic Interaction Networks, with Enno Mammen and Wolfgang Polonik (Accepted at Journal of Business and Economic Statistics) - R-Code
• Inference in Regression Discontinuity Designs with High-Dimensional Covariates, with Christoph Rothe, The Econometrics Journal, Volume 26, Issue 2, May 2023, Pages 105-123 - R-Code
• Semi-Parametric Estimation of Incubation and Generation Times by Means of Laguerre Polynomials, with Ingrid Van Keilegom, Journal of Nonparametric Statistics, 34:3, 570-606, 2022 - R-Code
• Correlation bounds, mixing and m-dependence under random time-varying network distances with an application to Cox-Processes, Bernoulli 27 (3) 1666 - 1694, August 2021 - R-Code
• Nonparametric inference for continuous-time event counting and link-based dynamic network models, with Enno Mammen and Wolfgang Polonik, Electron. J. Statist. 13 (2) 2764 - 2829, 2019 - R-Code

Other Publications

• Book Review:Skyler J. Cranmer, Bruce A. Desmarais, Jason W. Morgan: Inferential network analysis. Stat Papers 63, 989–990 (2022).

Programme-Code

• The R-package QuantRegLaguerre can be downloaded here from github. It provides implementations for the quantile regression methodology used in the paper Efficient Quantile Regression under Censoring Using Laguerre Polynomials written by Ingrid Van Keilegom and myself. It contains functions for computing the estimator, performing cross-validation, and finding the bootstrap estimate for the variance.
• The R-package HighDimRD which contains the methodology from Inference in Regression Discontinuity Designs with High-Dimensional Covariates can be downloaded from the github repository HighDimRD.
• The R-code used in the paper Semi-Parametric Estimation of Incubation and Generation Times by Means of Laguerre Polynomials is available on github in the repository SemiParametric-Laguerre.
• R and C-code which implements the procedures studied in my papers about counting processes on networks can be found on github in the repositories Estimate-Event-Network and Baseline-Estimation.

Online Seminar Talks

• German Probability and Statistics Days, October 2021: Website - Video
• EcoDep Seminary, April 2021: Website - Video
• Bernoulli-IMS One World Symposium, August 2020: Website - Video

Teaching

• Summerterm 2024 (Leipzig University): Mathematische Statistik (Further Information on Moodle)
• Winterterm 2023/24 (Leipzig University): Causal Inference in Statistics
• Summerterm 2023 (Leipzig University): Mathematische Statistik
• Winterterm 2022/23 (Leipzig University): Statistical Network Analysis
• Summerterm 2022 (Leipzig University): Mathematische Statistik
• Springterm 2019 (University of Mannheim): Econometrics of Networks