Data Mining Group – TU Dortmund

• Archetype Discovery from Taxonomies: A Method to Cluster Small Datasets of Categorical Data, Lars Lenssen, Philip Stahmann, Christian Janiesch, Erich Schubert, 58th Hawaii Int. Conf. on System Sciences (HICSS), to appear
• Hierarchical Clustering Without Pairwise Distances by Incremental Similarity Search, Erich Schubert,Int. Conf. on Similarity Search and Applications, SISAP, 2024
• Advancing the PAM Algorithm to Semi-supervised k-Medoids Clustering, Miriama Jánošová, Andreas Lang, Petra Budikova, Erich Schubert, Vlastislav Dohnal, Int. Conf. on Similarity Search and Applications, SISAP, 2024
• Grouping Sketches to Index High-Dimensional Data in a Resource-Limited Setting, Erik Thordsen, Erich Schubert, Int. Conf. on Similarity Search and Applications, SISAP, 2024

Our group received three awards in one week:

• best student paper award at SISAP 2023 for a paper improving sketch-based similarity search in high-dimensional data: Erik Thordsen and Erich Schubert. An Alternating Optimization Scheme for Binary Sketches for Cosine Similarity Search
• best paper award at SISAP 2023 for the same contribution
• best paper award of the SIG Knowledge Discovery, Data Mining and Machine Learning (KDML) at the LWDA 2023 for improving nearest-neighbor consistent clustering: Lars Lenssen, Niklas Strahmann and Erich Schubert. Fast k-Nearest-Neighbor-Consistent Clustering

The following new publications were accepted:

• Lars Lenssen and Erich Schubert. Medoid Silhouette clustering with automatic cluster number selection. In: Information Systems 120, 102290, 2024 (online 2023)
• Lars Lenssen, Niklas Strahmann and Erich Schubert. Fast k-Nearest-Neighbor-Consistent Clustering In: Proceedings of the Conference "Lernen, Wissen, Daten, Analysen" (LWDA), Marburg, Germany, 2023, to appear. best paper award
• Melanie Derksen et al. Who Did What When? Discovering Complex Historical Interrelations in Immersive Virtual Reality In: 22nd IEEE International Symposium on Mixed and Augmented Reality (ISMAR), 2023, to appear.

The following new publications were accepted:

• Erich Schubert. Stop using the elbow criterion for k-means and how to choose the number of clusters instead. In: ACM SIGKDD Explorations 25(1), 36–42, 2023
• Erik Thordsen and Erich Schubert. An Alternating Optimization Scheme for Binary Sketches for Cosine Similarity Search. In: Similarity Search and Applications (SISAP), 41-55, 2023, best paper & best student paper award
• Andreas Lang and Erich Schubert. Accelerating k-Means Clustering with Cover Trees. In: Similarity Search and Applications (SISAP), 148–162, 2023

In winter term 2023/2024 the following modules will be offered by the group:

• Lecture Natural language processing
• Lecture Industrial data science
• Project group Facts4Chat
• Study project AI for board games

The following classes are offered in summer 2023:

• Presentation course (for Proseminar)
• Proseminar: Hacker Culture, Hacker Spirit & Hacker Ethics [German]
• Seminar: Ähnlichkeitssuche in großen Datenmengen [German]
• Industrial Data Science 2

The Stanford statistician John Ioannidis publishes a list of the 100,000 most influential scientists (science-wide).

In the single-year ranking of the current version (based on the data for the year 2021, published November 2022), Prof. Schubert ranks 92735.

YouTube lecture recordings of Corona lectures are available in our YouTube channel.

New lectures will not be recorded.

Active participation in class leads to better learning outcomes.