Matthias Carnein


I am a 29-year-old PhD candidate at the University of Münster, Germany where I work at the Information Systems and Statistics Group. My research focuses on methods used in data science including the analysis of data streams and clustering with applications in market segmentation and customer relationship management.

Education

since 03/2016
PhD
PhD candidate in Information Systems at the University of Münster, Germany. Large parts of my PhD were carried out in collaboration with Arvato, a subsidiary of Bertelsmann.
10/2013 – 09/2015
Master of Science
Master Studies in Information Systems at the University of Münster, Germany
10/2010 – 09/2013
Bachelor of Science
Bachelor Studies in Information Systems at the University of Münster, Germany

Positions

since 03/2016
Research Assistant
Information Systems and Statistics Group at the University of Münster, Germany
04/2017 – 12/2017
Lecturer
Certificate programme ‘Data Science’ at the Münster University Continuing Education (WWU Weiterbildung)
01/2016 – 02/2016
Graduate Student Assistant
Information Systems and Statistics Group at the University of Münster, Germany
04/2014 – 12/2015
Graduate Student Assistant
IT Security Group at the University of Münster, Germany
11/2012 – 03/2014
Student Assistant
Information Systems and Information Management Group at the University of Münster, Germany

Conferences & Events

September 2019 European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECMLPKDD ’19) in Würzburg, Germany
July 2019 21st IEEE Conference on Business Informatics (CBI ’19) in Moscow, Russia
April 2019 23rd Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD ’19) in Macau, China
February 2019 Visiting scholar at the University of Waikato in Hamilton, New Zealand
January 2019 Visiting scholar at the Queensland University of Technology in Brisbane, Australia
July 2018 1st Metaheuristics Summer School (MESS) in Sicily, Italy
November 2017 36th International Conference on Conceptual Modeling (ER ’17) in Valencia, Spain
August 2017 German-Brazilian Workshop on Information Systems in Logistics and Production Engineering in Recife, Brazil
May 2017 ACM International Conference on Computing Frontiers (CF ’17) in Siena, Italy
March 2017 17th Conference on Database Systems for Business, Technology, and Web (BTW ’17) in Stuttgart, Germany
9th International Conference on Evolutionary Multi-Criterion Optimization (EMO ’17) in Münster, Germany
June 2016 11th Madrid UPM Advanced Statistics and Data Mining Summer School in Madrid, Spain
Februrary 2016
IS&T Electronic Imaging: Media Watermarking, Security, and Forensics (EI ’16) in San Francisco, USA
November 2015 7th IEEE International Workshop on Information Forensics and Security (WIFS ’15) in Rome, Italy
June 2014 2nd ACM Workshop on Information Hiding and Multimedia Security (IH & MMSec ’14) in Salzburg, Austria

Teaching

Summer Term 19
  • Auto-ML (Master Seminar)
  • Data Analytics II: Support Vector Machines (Master Lecture)
  • Dialogue Act Classification in the CRM Context (Master Thesis)
Winter Term 18/19
  • Introduction to Information Systems: Statistics / Analytics (Bachelor Lecture)
Summer Term 18
  • Text Mining to Prioritize Maintenance Tasks in Agile Software Projects (Master Thesis)
  • Identification of Disease Candidates based on Medical History (Master Thesis)
  • Data Analytics II: Support Vector Machines (Master Lecture)
  • Chat Bot for the Examination Office (Bachelor Project Seminar) [News & Press] [News & Press] [News & Press] [News & Press]
Winter Term 17/18
  • Customer Service at HILTI (Master Project Seminar)
  • Statistical Methods in Retail (Master Seminar)
  • Introduction to Information Systems: Statistics / Analytics (Bachelor Lecture)
  • Data Analytics I: R-Course (Master Lecture)
  • Image Segmentation for the Microstructural Analysis of Engine Materials (Master Thesis)
Summer Term 17
  • Data Analytics II: Support Vector Machines (Master Lecture)
Winter Term 16/17
  • Stream Clustering (Master Seminar)
  • Applied Machine Learning (Master Seminar)
  • Tech-enabled Omni-channel CRM (Bachelor Project Seminar)
  • Data Analytics I: Stream Clustering (Master Lecture)
  • Empirical Comparison of Stream Clustering Algorithms (Bachelor Thesis)
  • Applied Customer Base Analysis in a Noncontractual Setting Based on Transactional Data (Master Thesis)
Summer Term 16
  • Data Analytics II: Support Vector Machines (Master Lecture)

Tools & Software

MOA /
confStream
I am a contributor to the Massive Online Analysis (MOA) framework. MOA is the most popular framework for stream data mining. My contribution is currently maintained as a fork and implements the confStream algorithm as proposed in our ECMLPKDD ’19 paper.
stream I am a contributor to the popular R-Package ‘stream’ which implements various functions and algorithms to cluster data streams. The package was used for our ER ’17 paper, CF ’17 paper, BISE article and Big Data Research article.
streamMOA I am a contributor to the R-Package ‘streamMOA’ which interfaces the stream clustering algorithms available in the Massive Online Analysis (MOA) library. The package was used for our CF ’17 paper, BISE article and Big Data Research article.
evoStream An evolutionary stream clustering algorithm. The algorithm is able to utilize the idle time in a stream in order to incrementally improve the clustering result. The package was used for our Big Data Research article.
userStream Implementation of a stream clustering algorithm applicable to customer segmentation. The algorithm allows to identify and track segments of similar customers over time as proposed in our PAKDD ’19 paper.
textClust A stream clustering algorithm which can identify and track topics in a stream of texts as proposed in our ER ’17 paper.
Customer Service
Monitor
A tool to analyse customer service on Social Media. The tool allows to automatically determine how often and frequently companies respond to questions on facebook and twitter. The tool was used for our BTW ’17 paper.
jpegToolbox An R-Package that provides various implementations of the JPEG image compression algorithm. Most importantly, an interface to the popular libjpeg library is available. The package ships with pre-compiled libraries of libjpeg versions 6b, 8d and 9a for Windows. In addition, the entire lossy compression pipeline is implemented in R for easy debugging as well as in C++ with interfaces to R for faster computation speed. The implementations were used for our EI ’16 paper and WIFS ’15 paper.

Publications