Table 2.

Comparison of various cluster validity indices between our proposed method and k-means clustering for AD vs control.

Cluster validity index	Proposed method	K-means clustering
Ball_hall	0.331	0.111
Davies_bouldin	5.205	2.090
Dunn	0.082	0.009
G_plus	0.136	0.070
Gdi11	0.082	0.009
Gdi12	0.363	0.068
Gdi31	0.394	0.209
Ray_turi	14.018	2.229

Higher value signifies better than the other value in the same row (cluster validity index).