News

 Fri, 10. Jan. 2020   Beck, Fabian

Learning from Network Comparison

New interdisciplinary MERCUR research project on comparative data analysis within the Ruhr University Alliance

When evaluating their environment, humans tend to compare things with each other in order to learn from them. Here "things" are to be understood in a general way and are by no means limited to objects. For example, we compare our own behavior with that of others. In business, too, companies compare themselves with competitors in order to assess what they do well or where they can improve. But comparisons are not always easy. Especially when it comes to dynamic behavior, the temporal relationships and dependencies are often represented by very complex networks.

In a new interdisciplinary research project, researchers at the TU Dortmund University and the University of Duisburg-Essen want to develop new comparative data analysis methods. These methods are intended to show ways in which data from dynamic processes can be transferred into a uniform network model and compared as such. Project leader Prof. Carsten Jentsch (Statistics, TU Dortmund University) says: "The abstraction of complex data structures as dynamic networks allows novel access for statistical analysis methods." Jun-Prof. Fabian Beck (Data Visualization, University of Duisburg-Essen) adds: "We not only want to express the differences in numbers, but also understand the how and why. Therefore, statistics and data visualization work closely together in this project." In particular, they will focus on logistics, which is a key industry in the Ruhr region and is represented in the project by Jun.-Prof. Anne Meyer (Logistics, TU Dortmund). Futher, applications in the field of human-computer interaction and IT security, which are being researched at the Essen-based software technology institute paluno, are also to be investigated. Prof. Roland Fried (Statistics, TU Dortmund University) supports the detection of changes by means of comparative analyses in monitoring dynamic network processes.

The project "Comparative Analysis of Dynamic Network Structures Combining Statistical and Visual Methods" is supported by the Mercator Research Center Ruhr with a MERCUR project grant of approx. 400,000 Euros and runs for three years. The project also aims to strengthen cooperation within the University Alliance Ruhr and to launch a joint research agenda in this field.