> ML_LITERATURE // ESTER-1996-DBSCAN-DENSITY-BASED-SPATIAL-CLUSTERING_v1.0
A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise
Martin Ester, Hans-Peter Kriegel, Jörg Sander, Xiaowei Xu · ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD) (1996)
foundational1996foundationalthirdPartyReproduced
Principal Contribution
Formulated density-based clustering using epsilon-neighborhoods and MinPts, discovering arbitrary-shaped clusters and isolating noise.
Operational Relevance
Primary tool for geospatial clustering, GPS trajectory anomaly detection, and discovering non-spherical clusters.
Assumptions
- Clusters are contiguous regions of high point density separated by low-density noise regions
Limitations
- Fails when clusters have varying densities across the same dataset (resolved by HDBSCAN)
Connected Algorithms, Architectures & Tools
Related Algorithms:
Related Architectures:
Implementing Libraries: