Record link: 
http://hdl.handle.net/20.500.12708/231967
-
Title: 
On the Synthetic Evaluation of Time Series Clustering
en
Citation: 
Iglesias Vázquez, F., & Zseby, T. (2027). On the Synthetic Evaluation of Time Series Clustering. In Discovery Science : 29th International Conference, DS 2026, Mainz, Germany, October 5–9, 2026, Proceedings (pp. 479–494). Springer Cham. https://doi.org/10.1007/978-3-032-41086-3_31
-
Publisher DOI: 
10.1007/978-3-032-41086-3_31
-
Publication Type: 
Inproceedings - Full-Paper Contribution
en
Language: 
English
-
Authors: 
Iglesias Vázquez, Félix 
Zseby, Tanja 
-
Organisational Unit: 
E389-01 - Forschungsbereich Networks 
E056-10 - Fachbereich SecInt-Secure and Intelligent Human-Centric Digital Technologies 
E056-16 - Fachbereich SafeSeclab 
-
Published in: 
Discovery Science : 29th International Conference, DS 2026, Mainz, Germany, October 5–9, 2026, Proceedings 
-
ISBN: 
978-3-032-41085-6
-
Volume: 
17130
-
DOI of the book: 
10.1007/978-3-032-41086-3
-
Date (published): 
17-Oct-2027
-
Event name: 
29th International Conference Discovery Science (DS 2026) 
en
Event date: 
5-Oct-2026 - 9-Oct-2026
-
Event place: 
Mainz, Germany
-
Number of Pages: 
16
-
Publisher: 
Springer Cham
-
Peer reviewed: 
Yes
-
Keywords: 
Time series clustering; Synthetic benchmark generation; Machine learning evaluation
en
Abstract: 
High-quality evaluation remains a major challenge in machine learning, particularly in time series clustering. Recent benchmarks, including deep-learning approaches, report that no method significantly outperforms the decade-old k-Shape algorithm. This raises the question of whether progress in the field is genuine or limited by inadequate evaluation infrastructures.
en
Research Areas: 
Mathematical and Algorithmic Foundations: 40%
Information Systems Engineering: 20%
Computer Science Foundations: 40%
-
Science Branch: 
1020 - Informatik: 40%
2020 - Elektrotechnik, Elektronik, Informationstechnik: 20%
1010 - Mathematik: 40%
-
Appears in Collections:
Conference Paper

Show full item record

Google ScholarTM

Check