DBSCAN
- Date
- 1996 · confidence high
- Place
- Munich, Bavaria, Germany
- Credited to
- Martin Ester, Hans-Peter Kriegel, Jörg Sander and Xiaowei Xu
- Field
- Knowledge & symbols
What it is
A density-based clustering algorithm that identifies groups of closely packed points without requiring the number of clusters to be fixed in advance, marking isolated points as noise; proposed by Martin Ester, Hans-Peter Kriegel, Jörg Sander and Xiaowei Xu.
What it led to
- A reference standard in data mining and unsupervised learning
- Variants such as HDBSCAN and OPTICS
- SIGKDD Test of Time Award (2014)