Design and Implementation of an Improved K-Means Clustering Algorithm

Table 2

Test results of three algorithms on 10 datasets.

UCI data set	Evaluation index	K-means	Canopy kmeans	Article algorithm
Iris	s	0.62	0.68	0.78
	Acc%	71.34	80.22	86.56
	ARI	0.75	0.78	0.81
	NMI	0.73	0.79	0.83
	F-score	0.74	0.83	0.88
Wine	s	0.57	0.66	0.73
	Acc%	68.61	74.85	82.15
	ARI	0.65	0.72	0.78
	NMI	0.64	0.70	0.81
	F-score	0.66	0.71	0.79
Glass	s	0.51	0.58	0.64
	Acc%	60.14	68.65	73.62
	ARI	0.58	0.67	0.74
	NMI	0.62	0.65	0.77
	F-score	0.62	0.66	0.75
TAE	s	0.63	0.72	0.79
	Acc%	76.12	82.57	88.27
	ARI	0.72	0.79	0.85
	NMI	0.74	0.80	0.86
	F-score	0.75	0.81	0.87
Soybean-(small)	s	0.56	0.61	0.69
	Acc%	67.31	72.56	80.21
	ARI	0.66	0.70	0.76
	NMI	0.68	0.69	0.78
	F-score	0.69	0.74	0.82
Blood	s	0.59	0.68	0.77
	Acc%	70.32	77.16	86.12
	ARI	0.68	0.76	0.84
	NMI	0.68	0.78	0.85
	F-score	0.68	0.73	0.85
Diabetes	s	0.47	0.53	0.61
	Acc%	55.33	62.18	70.12
	ARI	0.56	0.60	0.68
	NMI	0.57	0.58	0.66
	F-score	0.56	0.64	0.72
Seeds	s	0.62	0.68	0.76
	Acc%	71.38	79.52	88.89
	ARI	0.67	0.75	0.83
	NMI	0.69	0.79	0.86
	F-score	0.67	0.76	0.85
Wave	s	0.42	0.43	0.51
	Acc%	52.29	55.63	61.32
	ARI	0.44	0.47	0.58
	NMI	0.46	0.48	0.56
	F-score	0.48	0.49	0.59
Ecoli	s	0.49	0.55	0.63
	Acc%	59.69	66.73	74.12
	ARI	0.56	0.62	0.71
	NMI	0.59	0.65	0.75
	F-score	0.61	0.64	0.73