Development of a Practical Theory for Clustering Algorihtms through Data-Driven Modeling and Analysis
Goals
By cluster analysis or simply clustering one understands the partitioning of a set of objects into subsets of similar objects. Clustering algorithms apply in data compression, pattern recognition, biology, analysis of networks, stochastics, text classification and machine learning, to give only a few examples. The varying applications also determine what is to be understood by similar objects. On the one hand there exists a quantity of different clustering algorithms successfully used in practice. On the other hand there is also a multiplicity of research results concerning clustering coming from theoretical computer science. However nearly all practically used algorithms can only be analyzed insufficiently and the algorithms derived from theory are not efficient enough for practical use. In this project we want to try and close this gap between theory and practice by a practice-oriented theory for clustering algorithms. Emphasis is to be placed on modelling and the arising algorithm analysis. It takes the characteristics of the input into consideration by a suitable parametrization.
The project ist part of the priority program "Algorithm Engineering" funded by the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG). The project starts with Christian Sohler and Johannes Blömer jointly in Paderborn and is managed now by Christian Sohler at the University of Dortmund and by Johannes Blömer at the University of Paderborn.
Members
- Johannes Blömer (Project Manager)
- Christian Sohler (Project Manager, formerly University of Paderborn, currently University of Dortmund)
- Daniel Kuntze