At the end of last year, we had the pleasure of presenting our publication of Lennart Purucker‘s Master thesis at the 22nd IEEE International Conference on Data Mining (ICDM) in Florida, USA!
Lennart worked on Meta-Learning for Subgroup Discovery during his Master’s thesis at the RWTH Aachen University. As part of this thesis, a novel estimation method for the pruned search space size of Subgroup Discovery was found.
After Lennart joined the ISG Siegen, he and Prof. Joeran Beel worked with Felix Stamm (Ph.D. at the RWTH Aachen University) and Prof. Florian Lemmerich (Professor at the University of Passau) to publish the novel estimation method.
Subgroup Discovery is a pattern mining approach that can take a long time because the search space can increase exponentially depending on the mining parameters. Our method estimates how large the pruned search space is so that practitioners can adjust the parameters if needed. Both the code and the paper are now publicly available.