Fig. 5: DBOpt performance on 3D simulated data. | Communications Biology

Fig. 5: DBOpt performance on 3D simulated data.

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

Fig. 5

a Representative simulation (S01, Table S1f) of 3D clusters, b optimal clustering of the representative simulated data via DBOpt, c comparisons of V-measure scores between naive EVS and V-measure scores for parameters chosen with DBOpt for both ellipsoid and fibrillar clusters, d distributions of the ratio of DBOpt V-measure scores to maximum naive EVS V-measure scores (dotted line represents median and “X” represents mean; n = 20 simulations), and e maximum V-measure and f maximum DBCV scores for each algorithm found via naive EVS and DBOpt, respectively (n = 20 simulations). Box indicates 25th and 75th percentile, whiskers indicate 5th and 95th percentile, and the interior line indicates median. Statistical analysis of d–f is discussed in Supporting Text 4, Table S2d.

Back to article page