Fig. 3: Naive EVS and DBOpt parameter sweeps. | Communications Biology

Fig. 3: Naive EVS and DBOpt parameter sweeps.

From: Bayesian optimized parameter selection for density-based clustering applied to single molecule localization microscopy

Fig. 3

Elliptic (E14), micellular (M22), and mixed (V12) parameter sweeps of cluster scenarios are also shown in Fig. 2 for DBSCAN, HDBSCAN, and OPTICS, with optimal EVS and DBOpt cluster outcomes plotted in the right panel. Parenthetical notations indicate the simulated data as described in Table S1. Points on the contour plots represent each parameter combination that is scored from which the contour plot is prepared (discrete sampling for naive EVS, Bayesian sampling for DBOpt); the white “X” represents the optimal chosen parameters from EVS and DBOpt. Right panel plots show clusters identified from the highest scoring set of parameters from the highest scoring algorithm and correspond to those marked with * in Fig. S4.

Back to article page