Fig. 3.

 (A) and (B): Average performance values with one standard deviation of the six clustering methods when γ = 0.01 (A) and γ = 0.006 (B). The proportions of missing values are 37% and 90% for (A) and (B), respectively. The results are based on simulating 50 datasets following Simulation model 1; C and D: Average performance values with one standard deviation of the seven clustering methods when γ = 0.6 (A) and γ = 0.1 (B). The proportions of missing values are 56 and 67% for (C) and (D), respectively. The results are based on 50 datasets simulated following Simulation model 2. See Section E of the Supplementary Material for details of the simulation models and γ