Label Propagation Clustering Algorithm Based on Adaptive Angle

Table 6

Comparison of clustering effects on the UCI datasets.

Dataset	Criteria	DBSCAN	FCM	DBC	ALR-DBC
Iris	NMI	0.7336	0.7433	0.8705	0.8980
	ARI	0.5681	0.7287	0.8857	0.9221
	Homogeneity	0.5793	0.7404	0.8696	0.8980
Dermatology	NMI	0.6205	0.7644	0.8102	0.8486
	ARI	0.4151	0.6866	0.7307	0.7871
	Homogeneity	0.6484	0.6878	0.7859	0.7879
Balance	NMI	0.0178	0.0093	0.2082	0.2104
	ARI	0.0200	0.0075	0.0424	0.0440
	Homogeneity	0.0101	0.0083	0.5125	0.5181
Vote	NMI	0.3977	0.4547	0.4942	0.4942
	ARI	0.4480	0.5232	0.5709	0.5709
	Homogeneity	0.5034	0.4633	0.5030	0.5030
Vowel	NMI	0.5317	0.5420	0.5576	0.5693
	ARI	0.4170	0.4004	0.4221	0.4579
	Homogeneity	0.4902	0.5503	0.5242	0.5199
Landsat	NMI	0.5768	0.6054	0.6415	0.6365
	ARI	0.4153	0.5216	0.5966	0.6306
	Homogeneity	0.5026	0.6091	0.6667	0.6395
Ecoli	NMI	0.6161	0.5448	0.6550	0.6571
	ARI	0.1459	0.3620	0.7009	0.7165
	Homogeneity	0.6285	0.6431	0.5558	0.5580
WDBC	NMI	0.3790	0.6151	0.5203	0.5258
	ARI	0.4661	0.7304	0.5189	0.5883
	Homogeneity	0.3868	0.6080	0.6528	0.6656