Grids showing the evaluation of the 2025 sample’s DBSCAN classification. The evaluation was performed in ρ_ECM-log10(mass) coordinates with different MinPts and epsilon values. The evaluation is a collection of bootstrapping runs with 200 samples. For each bootstrap sample, two grids of evaluation scores are made by running the DBSCAN on the bootstrapped sample over a range of 200 epsilon values from 0.1 to 0.4 and MinPts from 3 to 20. One grid records the silhouette score, and one grid records the CHI. These two grids are used to find the optimal parameters for the 2025 dataset. In panel A, the total grid of all the recorded silhouette scores. In panel B, the total grid of all the recorded CHI. In both of the top panels, the best-performing epsilon value is marked with a red x.