Exploratory Data Analysis

Vision and Research Strategy

Thanks to modern technology, collecting and storing massive amounts of complex information in databases has become fairly straightforward. Examples include astronomical observations, records of hospital patients, as well as experimental data in the life-sciences. The complexity and the sheer volume of this data has led to the situation where analysis of the data---the actual goal of collecting and storing all this data---has become increasingly difficult, if not sheer impossible.
Our group's research goal is to solve this problem. That is, our research is inside the field of data mining and machine learning, with a focus on exploratory data analysis. We aim to develop theory and methods by which we are able to efficiently identify the most interesting and useful sub-structures in data of any size or shape. To this end we often employ insights from Information Theory and other well-founded statistical methods. The Minimum Description Length (MDL) and Maximum Entropy (MaxEnt) principles have proven to be valuable tools in particular.

Currently we are investigating statistical and information theoretic techniques for identifying informative local structures such as patterns in large graphs as well as large collections of real-valued data, how to efficiently mine good data descriptions directly from rich data, and study well-founded approaches for meaningfully comparing between, and validation of, exploratory data analysis results. 

Composition of Group

Our group was founded in October 2013 in the Cluster of Excellence. Currently it consists of dr. Jilles Vreeken (head), dr. Mario Boley (postdoc), Kailash Budhathoki, Janis Kalofolias, Panagiotis Mandros, Alexander Marx  (PhD students), and Amir Baradaran, Iva Baykova, Robin Burghartz, Maike Eissfeller, Patrick Ferber, Jonas Fisscher, Xinguang Gao, Magnus Halbe, Michael Hedderich, and Frauke Hinrichs (research assistants).

We are always looking for PhD candidates, postdocs, or HiWis, with background and interest in data mining, machine learning, statistics, and/or mathematics.

Projects and Collaborations

Dr. Jilles Vreeken

Dr. Jilles Vreeken

Jilles Vreeken is the head of the Independent Research Group on Exploratory Data Analysis within the Cluster of Excellence.

Fon: +49 681 302 71925