Research

Projects funded by the German Research Foundation

Since 2004, Prof Schmid has been a successful applicant for nine DFG projects with a total volume of around 1.6 million euros.

Funding/scope

DFG / 223,950 &euros;

Term

2023-2026

Cooperation partner

  • Prof. Dr Sven Knoth, Helmut Schmidt University
  • Prof. Dr Yarema Okhrin, University of Augsburg

 
Description

The aim of the project is to develop statistical methods for the rapid detection and identification of changes in matrix-valued processes. Such process structures can be found today in many scientific disciplines such as engineering, economics and information science. The rapid detection and identification of changes is of great relevance, for example to reduce production costs and optimise process sequences, to detect environmental changes at an early stage and to quickly uncover suspicious activities in networks. The new methods developed in the theoretical part of the project are based on the further development of sequential monitoring methods for vector-valued data to the hitherto little explored area of matrix-valued data sources. We focus on the construction of tools and methods that capture the complex
temporal dependencies and cross-sectional dependencies between the components of the matrix-valued data stream. Particular attention is paid to the scalability of the methods in order to ensure applicability to high-dimensional data sources. In addition, the associated computational problems for calculating the performance measures of the monitoring methods are analysed in detail. The proposed methods have a wide range of applications covering all areas characterised by matrix-valued data. In the application part of the project, we focus on important examples from different areas of social and natural sciences, but this does not by far exhaust the possibilities. We study financial risks quantified by covariance matrices, image processes given by the matrices of pixel intensities and networks characterised by adjacency matrices.

Teaching
 

Publications
 

Project management and employees

  • Prof. Dr Wolfgang Schmid
  • Viktoriia Petruk

 

Funding/scope

DFG / 9,200 €

Term

2019 - 2022

Cooperation partner


Description

We are planning a series of theoretical and applied subprojects in the field of spatial ARCH/GARCH and other volatility models. Firstly, the class of spatial ARCH processes will be extended to spatial GARCH models and exponential spatial GARCH models. Furthermore, the second point will focus on the development of multivariate ARCH processes. The spatial ARCH models introduced are also suitable for modelling spatiotemporal processes (cf. Otto, Schmid and Garthoff 2017). For empirical questions, a time-dependent spatial dependency structure is often observed, which can be represented, for example, by time-varying weighting matrices. Such weighting matrices often depend on stochastic variables such as migration intensity or wind speed and direction. Consequently, the assumed dependency structure is no longer deterministic and therefore, thirdly, the random influences on the probability structure and parameter estimation of the process should be analysed. Fourthly, sequential monitoring methods for spatio-temporal ARCH processes are to be developed in order to detect changes in the spatial dependency structure, for example.
As the tasks of the four broad fields of work are closely interlinked and interrelated, we would first like to explain the joint work programme of the two applicants and then present the specific tasks and fields of work separately for each applicant.

Teaching
 
Publications
 
Project management and employees
  • Prof. Dr Wolfgang Schmid

Funding / scope

DFG / 189,700 &euros;

Term

2020 - 2023

Cooperation partner

  • Prof. Dr Philipp Otto, Leibniz Universität Hannover

 

Description

The rapid development of computer technology in recent years has made it possible to store and analyse huge data sets. The dimension of the underlying stochastic model is very large and new approaches have to be developed in order to draw statistical conclusions for such a situation. If the dimension of the data is large, the classical limit values can no longer be applied directly and the classical estimators can asymptotically deviate drastically from the estimated parameters. The subject of this project is the monitoring of high-dimensional processes. This is a completely new discipline. The majority of previous publications in this area are based on independent random processes. This is a drastic limitation that is often not fulfilled in practice. It is assumed here that the underlying process is a high-dimensional time series and that the data therefore has a dependency structure. The aim is to detect a change in the position or covariance behaviour as soon as it occurs. For this purpose, control methods from the field of multivariate statistical process control must be extended and adapted to high-dimensional data. Within the scope of the project, different types of new control charts are to be derived by applying the (generalised) likelihood-ratio method and the (modified) Shiryaev-Roberts method, among others. A distinction is made between control charts for the positional behaviour and charts for the covariance structure and simultaneous control procedures are to be introduced. They detect changes in the position or covariance behaviour. The introduced control schemes will be compared using different quality criteria such as average run length and average delay. Various applications are discussed in this project. Such problems occur in industrial manufacturing as a consequence of new digital production methods in the context of Industry 4.0. New technologies (additive manufacturing, micro manufacturing,..) in combination with new testing methods (contactless systems, X-ray computed tomography,..) and fast multi-streaming high-speed sensors (acoustic, temperature, pressure signals,....) are paving the way for a new generation of industrial big data, the analysis of which requires new modelling approaches and monitoring methods to ensure error-free production. Another important area of application is the monitoring of portfolios. A portfolio is often made up of a large number of shares and the number is large in relation to the number of shares being monitored. For an investor, it is now important to recognise a change in risk behaviour at an early stage in order to be able to rebalance their portfolio in good time.

Teaching
 
Publications
 
Project management and employees

  • Prof. Dr Wolfgang Schmid
  • Diana Ivaniuk
  • Rostyslav Bodnar

Funding / Scope
DFG / 183,550 €

Term
2023 - 2026

Cooperation partner
Prof. Dr Sven Knoth, University of the Federal Armed Forces Hamburg

Prof. Dr Yarema Okhrin, University of Augsburg

Description

The aim of the project is to develop statistical methods for the rapid detection and identification of changes in matrix-valued processes. Such process structures can be found today in many scientific disciplines such as engineering, economics and information science. The rapid detection and identification of changes is of great relevance, for example to reduce production costs and optimise processes, to detect environmental changes at an early stage and to quickly uncover suspicious activities in networks. The new methods developed in the theoretical part of the project are based on the further development of sequential monitoring methods for vector-valued data to the hitherto under-researched area of matrix-valued data sources. We focus on the construction of tools and methods that capture the complex temporal dependencies and cross-sectional dependencies between the components of the matrix-valued data stream. Particular attention is paid to the scalability of the methods to ensure applicability to high-dimensional data sources. In addition, the associated computational problems for calculating the performance measures of the monitoring methods are analysed in detail. The proposed
methods have a wide range of applications covering all areas characterised by matrix-valued data.
In the application part of the project, we focus on important examples from different areas of social and natural sciences, but this by far does not exhaust the possibilities. We study financial risks quantified by covariance matrices, image processes given by the matrices of pixel intensities and networks characterised by adjacency matrices.

Teaching

Publications

Project management and staff

Prof. Dr Wolfgang Schmid

Viktoriia Petruk

Qualifications

Prof Schmid has been the first supervisor of 28 dissertations to date. Two of the dissertations took place at the University of Ulm and 23 dissertations at the European University Viadrina. There have also been five habilitations at his chair.
Two of his doctoral students, Taras Bodnar and Yarema Okhrin, were honoured with the Young Talent Award of the State of Brandenburg
in 2008 (endowed with 20,000 euros).
Twelve of his former doctoral students and staff are now professors at universities in Germany, England, the Netherlands, Sweden, Scotland and Ukraine as well as at a German university of applied sciences.

Ongoing PhDs

Anna-Liesa Otto Binäre predictions based on complex data structures alange@europa-uni.de
Viktoriia Petruk Monitoring Image Processes petruk@europa-uni.de
Diana Ivaniuk Monitoring High Dimensional Time Series ivaniuk@europa-uni.de
Johannes Schwenzer Forecasting Toolbox for a climate-resilient unniversal electrification schwenzer@europa-uni.de

 

Completed PhDs

Dr Dmytro Ivasiuk 23 February 2022 New Results on Optimal Portfolio Selection for the Power Utility Function
Dr Ivan Semeniuk  09.02.2022 New Approaches for Monitoring Multivariate and Image Processes
Dr Solomiia Dmytriv  23.07.2020 Statistical Theory of High-Dimensional Portfolios
Dr Patrick Vetter  10.04.2017 Space-time modelling of the global CO2 concentration
Dr Liubov Rabyk 30 January 2017 Design of EWMA Control Schemes for Long-Memory Processes
Prof. Dr Daniel Ambach 30 November 2016 Benefits and opportunities of wind energy for Poland and Germany - economic and political aspects
Prof. Dr Philipp Otto 17 November 2016 Spatial and Spatiotemporal Stochastic Processes - Modelling and Detection of Spatial Structural Breaks with Applications in Economics and Biometrics
Dr Taras Lazariv 02.07.2015 02/07/2015
Prof. Dr Nestor Parolya 18 December 2013 Multi-Period and High-Dimensional Portfolio selection Problems
Dr. habil. Robert Garthoff 17 April 2013 Monitoring Multivariate Nonliner Time Series
Dr Sergiy Ragulin 14/02/2012 Sequential Control Procedures in Portfolio Management
Dr Iryna Okhrin 13 July 2010 Surveillance of the optimal portfolio composition
Prof. Dr Ostap Okhrin 06 February 2008 Hierarchical Archimedian Copulas: Structure Determination, Properties, Applications
Prof. Dr Taras Zabolotskyy 19.09.2007 Properties of Optimal Portfolio Weights and Their Characteristics
Prof. Dr Roman Kozhan 09.03.2006 Portfolio Selection under the Choquet Expected Utility Model
Dr Przemysław Śliwa  20.10.2005  Continuous Inspection Schemes for Monitoring Multivariate Linear and Nonlinear Time Series
Prof. Dr Olha Bodnar  05.07.2005 CUSUM Control Charts for Multivariate Financial Time Series 
Dr Dobromir Tzotchev  02.06.2005  Sequential Surveillance of Continuous-Time Interest Rate Models
Prof. Dr Taras Bodnar  08.11.2004  Optimal Portfolios in an Elliptical Model- Statistical Analysis and Tests for Efficiency
Prof. Dr Yarema Okhrin  09.07.2004 Distributional Properties and Estimation of Optimal Portfolios
Prof. Dr Vasyl Golosnoy   26.05.2004  Sequential Control of Portfolio Weights
Dr Maciej Rosolowski  01.10.2003   Control Charts for the Mean and the Autocovariances of Financial Time Series
Dr Bernd Schmid 11/07/2001 Pricing Credit Linked Financial Instruments - Theory and Empirical Evidence
Dr Stefan Schipper 09 November 2000 Sequential Methods for Detecting Changes in the Volatility of Economic Time Series
Prof. Dr Thomas Severin 10 June 1999 Sequential methods for detecting changes in the expected value structure of financial time series

 

PhDs at the University of Ulm

Dr Holger Kramer 18 June 1997 On Control Charts for Time Series
Dr Alexander Schöne 11/07/1997 New developments in statistical process control
Dr Thomas Flak 07/07/1992 Outlier tests based on extreme sums and methods for identifying outliers in time series

Dr. habil. Robert Garthoff 2021 Analysis, Prediction and Statistical Surveillance of Time Series and Spatiotemporal Processes  
Prof. Dr Olha Bodnar 2010 Risk management in the financial sector http://www.ptb.de/cms/fachabteilungen/abt8/fb-84/ag-842/mitarbeiter-842.html
Prof. Dr Taras Bodnar 2009 Statistical methods in the environmental sciences Associate Professor Department of Mathematics, Universität Stockholm
Prof. Dr Yarema Okhrin 2008 Modelling of stock returns Chair of Statistics, University of Augsburg
Prof. Dr Sven Knoth 2003 Statistical Process Control Chair of Computational Statistics, University of the Federal Armed Forces Hamburg

 

Prof. Schmid was involved in three dissertation procedures abroad, namely in 2002 with Prof Dr Manuel Morais at the TU Lisbon, in 2004 with Dr David Bock at the University of Göteborg and in 2017 with Dr Deliang Dai at the University of Vaxjo. He was also a member of Prof Dr Antonio Pacheco's habilitation committee at the Technical University of Lisbon in 2009.

Awards, achievements, merits

  • Elected Chairman of the German Statistical Society (DStatG, see http://www.dstatg.de ) in 2012, re-elected in 2016
  • Elected Member of the International Statistical Institute in 2002

  • 2023
    Invited speaker at the "XIVth International Workshop on Intelligent Statistical Quality Control" in Washington D.C. in August
    2023
  • 2022
    Invited speaker at the "Joint Statistical Meeting" in Washington D.C. in August 2022
  • 2020
    Invited speaker at the "Florida Meeting of the American Statistical Association" in Pensacola, USA in February 2020
  • 2019
    Invited speaker of the section "Finance and Statistics" at the "XXIVth Congress of the Portuguese Statistical Society" in Amarante, Portugal in November 2019
    Invited speaker at the "XIIIth International Workshop on Intelligent Statistical Quality Control" in Hong Kong in August 2019
    Invited speaker at the "Workshop on the Future of Statistics" in Berlin in July 2019
    Plenary lecture at the "International Workshop on Sequential Methods" in New York, USA in May 2019
  • 2018
    Invited speaker of the section "Dynamic Models and Structural Changes" at the "11th International Conference on Computational and Methodological Statistics" in Pisa, Italy in December 2018
    Invited talk at the "MAM Workshop in Probability Theory, Mathematical Statistics and Applications" in Västras, Sweden in October 2018
    Invited lecture at the "International Congress of Mathematicians" in Rio de Janeiro in August 2018
    Research stay at the University of Buenos Aires, Argentina in July 2018
    Plenary lecture at "JOCLAD (Jornadas de Classificacao a Analise de Dados) Meeting" in Lisbon in March 2018
  • 2017
    Research stay at CEMAT (Center for Computational and Stochastic Mathematics) in Lisbon, Portugal in October, 2018
    Invited talk in the section "Spatio-Temporal Models" at the "Conference on Climate and Environment" in Bergamo, Italy in July, 2017
    Invited talk at the "Quality and Productivity Research Conference" in Storrs, USA in June 2017
  • 2016
    Invited lecture at the "XIIth International Workshop on Intelligent Statistical Quality Control" in Hamburg in August 2016
    Invited lecture at the "Workshop on Spatial Statistics" in Dresden in June 2016
    Guest professor at the University of Bergamo, Italy from April to May 2015 - English-language event on the topic
    "Industrial Statistics" in the "Management, Information and Production Engineering" department
  • 2015
    Research stay at the University of Stockholm, Sweden in November 2015
    Invited speaker of the section "Change Point Analysis" at the Statistical Week in Hamburg in September 2015
    Invited speaker at the workshop "International Symposium on Statistical Process Monitoring" in Padua, Italy in July 2015
    Guest professor at the University of Bergamo, Italy from April to May 2015 - English-language course on "Industrial Statistics"  in the "Management, Information and Production Engineering" department
  • 2014
    Research stay at CEMAT (Center for Computational and Stochastic Mathematics) in Lisbon, Portugal in November 2014
    Guest professor at the University of Bergamo, Italy from March to April 2014 - English-language course on "Industrial Statistics" in the "Management, Information and Production Engineering" department Discussant at the "Second Stu Hunter Research Conference" in Tempe, USA in March 2014
  • 2013
    Invited lecture at the "XIth International Workshop on Intelligent Statistical Quality Control" in Sydney, Australia in August
    2013
    Invited lecture at the "International Symposium on Statistical Process Control" in Piraeus, Greece in July 2013
    Guest professor at the University of Bergamo, Italy from March to April 2013 - English-language event on the topic
    "Industrial Statistics" in the Department of "Management, Information and Production Engineering"
  • 2012
    Invited lecture at the workshop "Time Series: Models, Breaks and Applications" in Karlsruhe in February 2012
  • 2011
    Plenary lecture at the "Annual Meeting of the Portuguese Statistical Society" in Nazare, Portugal in September 2011
    Invited lecture at the "Third International Workshop in Sequential Methodologies" at Stanford University, USA in June 2011
  • 2010
    Invited lecture at the "Xth InternationalWorkshop on Intelligent Statistical Quality Control" in Seattle, USA in August 2010
    Invited lecture at the "International Symposium on Business and Industrial Statistics" in Portoroz, Slovenia in July 2010

1. Research activities
Prof. Schmid is the author of more than 170 publications in international journals (see https : ==www:researchgate:net=profile=Wolfgang Schmid2 ). Several of his works have appeared in A+-ranked journals.

2. Professional activities in the field of statistics

2.1 Activities in statistical societies
Prof. Schmid has held or has held leading positions in statistical societies for many years.

  • Chairman of the German Statistical Society (DStatG, 2012 - 2020)
  • Deputy Chairman of the DStatG (2020 - 2024)
  • Deputy Chairman of the Dachverband Statistik (2016 - 2022)
  • Member of the Executive Board of the German Statistical Society and Junior Staff Representative of the DStatG (2008-2012)
  • Chairman of the working group Stochastic Models for Reliability, Quality and Safety (2009-2013)
  • Chairman of the Committee Statistics in Science and Technology of the DStatG (1998-2002)
  • Deputy Chairman of the Committee Statistics in Science and Technology of the DStatG (1994-1998)

2.2 Membership of editorial boards and publication of conference proceedings
Prof. Schmid is co-editor of three international statistics journals and five conference proceedings on statistical quality control.

  • Associate Editor of the journal Sequential Analysis (since 2004)
  • Member of the editorial board of the journal AStA - Advances in Statistical Analysis (since 2005)
  • Associate Editor of the Journal of Multivariate Analysis (since 2018)
  • Co-editor of the following books:
    Frontiers in Statistical Quality Control 12 (with S. Knoth), Springer, 2021.
    Frontiers in Statistical Quality Control 12 (with S. Knoth), Springer, 2018.
    Frontiers in Statistical Quality Control 11 (with S. Knoth), Springer, 2015.
    Frontiers in Statistical Quality Control 10 (with H.-J. Lenz and P.-Th. Wilrich), Physica-Verlag.
    Frontiers in Statistical Quality Control 9 (with H.-J. Lenz and P.-Th. Wilrich): Physica-Verlag, 2010.

2.3 Organisation of international workshops and conferences since 2010
Prof. Schmid has co-organised and chaired the programme committee of a large number of conferences in Germany and abroad in recent years, including

  • Co-organiser (with S. Knoth) of the "14th Workshop on Intelligent Statistical Quality Control" in Washington D.C. in
    August 2023
  • Chair of the programme committee of the "Statistical Week" in Dresden in September 2020
  • Chair of the programme committee of the "Statistical Week" in Trier in September 2019 (approx. 500 participants)
  • Co-organiser (with S. Knoth) of the "13th Workshop on Intelligent Statistical Quality Control" in Hong Kong in August
    2019
  • Organisation of a section on "New Developments in Statistical Process Control" at the "International Workshop on
    Sequential Methods"
  • Chair of the programme committee of the "DAGStat Conference (Statistics under one roof)" in March 2019 in Munich (approx.
    800 participants
    • )
  • Chairman of the programme committee of the "Statistical Week" in Linz in September 2018 (approx. 500 participants)
  • Chairman of the programme committee of the "Statistical Week" in Rostock in September 2017 (approx. 500 participants)
  • Chairman of the programme committee of the "Statistical Week" in Augsburg in September 2016 (approx. 500 participants)
  • Co-organiser (with S. Knoth) of the "12th Workshop on Intelligent Statistical Quality Control" in Hamburg in August
    2016
  • Chair of the programme committee of the "DAGStat Conference (Statistics under one roof)" in March 2016 in Göttingen
    (approx. 700 participants)
  • Chairman of the programme committee of the "Statistical Week" in Hamburg in September 2015 (approx. 500 participants)
  • Organiser of the section "Analysis of Spatio-Temporal Data" of the conference of the "Italian Research Group for Environmental
    Statistics" in Bari, Italy in June 2015
  • Chairman of the programme committee of the "Statistical Week"  in Hanover in September 2014 (approx. 500 participants)
  • Chairman of the programme committee of the "Statistical Week" in Vienna, Austria in September 2013 (approx. 500 participants)
    • .
  • Co-organiser (with H.-J. Lenz and P.-Th. Wilrich) of the "11th Workshop on Intelligent Statistical Quality Control" in Sydney,
    Australia in August 2013
  • Organiser of a section on "Statistics and Econometrics" at the "German-Polish Joint Conference on Probability Theory
    and Mathematical Statistics" in Torun, Poland in June 2013
  • Chair of the programme committee of the "DAGStat Conference (Statistics under one roof)" in Freiburg in March 2013 (approx.
    700 participants)
  • Organiser of a section on "Statistical Process Control" at the "International Workshop in Sequential Methodologies"
    at Stanford University, USA in June 2011
  • Co-organiser (with H.-J. Lenz and P.-Th. Wilrich) of the "10th Workshop on Intelligent Statistical Quality Control" in Seattle,
    USA in August 2010

 

  • Research activities

1. In his publications, Prof Schmid provided fundamental results for the evaluation of control methods for time-dependent processes.

2. Prof. Schmid recognised the importance of statistical process control for new areas as early as the late 1990s and was one of the first to deal with the application of sequential monitoring methods in finance.

3. Prof. Schmid was also a pioneer in analysing the influence of parameter estimation on the characteristics of a portfolio. He has been working on this topic since the beginning of the last decade. His findings provide a completely new foundation for evaluating portfolios.

4. In his work on spatial statistics, he and his co-authors have applied the GARCH model of Engle (1982) and Bollerslev (1986) to spatial processes and space-time processes.

  • Institutional activities in the field of statistics

Prof. Schmid has held leading positions in statistical societies since 1994 and has made a significant contribution to the development of statistics in Germany during these years.

From 2012 - 2020 he was Chairman of the German Statistical Society (DStatG), since 2020 he has been Deputy Chairman. The DStatG was founded in 1911 and is the largest statistical society in Germany.

In his time in office to date, Prof Schmid has focussed on the following:

1. professionalisation of the office: The office is managed by a full-time employee.

2. strengthening the company's external image through a new internet presence, regular press releases, use of new media, etc.

3. expansion of the annual conference into an international conference

4. intensifying the promotion of young talent through financial support for foreign conference participation for young members and the awarding of travel grants to the society's annual conference

  • Promotion of young talent

Prof. Schmid has supported many young researchers through his work, both directly as a supervisor of dissertations (28 to date as first supervisor) and through his work for the DStatG. Almost half of his former doctoral students have gone on to pursue a university career. From 2008 to 2012, he was the DStatG's young researchers' representative and responsible for organising the young researchers' workshop. Around 30 doctoral students take part in this workshop every year, presenting their research topic to experienced academics in a small group.

Prof Dr Wolfgang Schmid

Secretariat Kerstin Zirkelbach

Office hours

by arrangement