Dbscan Custom Metric

cxcscmu/skilllearnbench-dbscan-parameter-tuning-dbscan-custom-metric

> How to run DBSCAN clustering with a custom anisotropic distance metric using sklearn and scipy. Use this skill whenever the user needs to cluster spatial points with a weighted or shape-adjusted distance, such as attenuating x vs y distances differently, or when DBSCAN needs a non-Euclidean metric defined as a callable. Covers cluster centroid computation and noise-point handling.