Table 3. Performances of clustering (purity).

Dataset	GNG	NG	SOM	k-means	GNG using no topology	SC
Blobs	0.9744	0.9632	0.4110	0.9690	0.9813	0.9671
Circles	1.0000	1.0000	0.5403	0.9997	1.0000	1.0000
Moons	0.9992	0.9884	0.5810	0.9328	0.9985	0.9933
Iris	0.5840	0.5648	0.5020	0.8473	0.8427	0.8533
Wine	0.4650	0.4649	0.4379	0.6656	0.6827	0.6742
Spam	0.7676	0.6063	0.6082	0.6095	0.7464	0.6070
CNAE-9	0.6711	0.5920	0.2913	0.4887	0.5706	0.1871
Digits	0.8572	0.8129	0.3356	0.7641	0.8025	0.6023
MNIST	0.6100	0.5801	0.2784	0.6754	0.5888	nan

Note:

GNG, NG, SOM, k-means, GNG using no topology, and SC mean ACSs with GNG, NG, SOM, k-means, and GNG using no topology, and spectral clustering, respectively. The purities are the mean of 100 runs with random initial values. The best purities are bold.