Skip to Main Content
Table 1:
Convergence Statistics for Selected Model Orders , 60, 100.
SXSVIXIS
K = 8 Parcellation NMI 0.672 0.08 0.681 0.08 0.693 0.07 0.657 0.04 
 Convergence (zkv3.5 0.8 5.6 1.6 5.3 1.5 6.1 1.6 
 Convergence (likelihood) 6.1 0.3 14.1 3.5 21.31 6.7 21.3 8.7 
K = 60 Parcellation NMI 0.732 0.01 0.738 0.04 0.737 0.04 0.732 0.01 
 Convergence (zkv3.9 0.8 5.1 1.3 5.0 1.3 5.7 1.3 
 Convergence (likelihood) 7.3 0.9 25.4 11.2 15.5 5.5 32.5 9.3 
K = 100 Parcellation NMI 0.735 0.01 0.743 0.04 0.741 0.04 0.741 0.04 
 Convergence (zkv3.9 0.6 4.2 1.6 4.7 1.5 5.6 1.2 
 Convergence (likelihood) 8.2 1.2 24.4 11.3 12.4 2.3 26.8 7.3 
SXSVIXIS
K = 8 Parcellation NMI 0.672 0.08 0.681 0.08 0.693 0.07 0.657 0.04 
 Convergence (zkv3.5 0.8 5.6 1.6 5.3 1.5 6.1 1.6 
 Convergence (likelihood) 6.1 0.3 14.1 3.5 21.31 6.7 21.3 8.7 
K = 60 Parcellation NMI 0.732 0.01 0.738 0.04 0.737 0.04 0.732 0.01 
 Convergence (zkv3.9 0.8 5.1 1.3 5.0 1.3 5.7 1.3 
 Convergence (likelihood) 7.3 0.9 25.4 11.2 15.5 5.5 32.5 9.3 
K = 100 Parcellation NMI 0.735 0.01 0.743 0.04 0.741 0.04 0.741 0.04 
 Convergence (zkv3.9 0.6 4.2 1.6 4.7 1.5 5.6 1.2 
 Convergence (likelihood) 8.2 1.2 24.4 11.3 12.4 2.3 26.8 7.3 

Notes: Parcellation NMI is the mean pairwise NMI between all 10 clustering runs: Convergence (zkv) is the mean number of iterations before cluster assignments do not change. Convergence (likelihood) is the mean number of iterations before change in log likelihood is less than 10. Results are shown standard deviation.

Close Modal

or Create an Account

Close Modal
Close Modal