. Author manuscript; available in PMC: 2018 May 1.

Published in final edited form as: IEEE Trans Neural Netw Learn Syst. 2016 Feb 18;28(5):1123–1138. doi: 10.1109/TNNLS.2015.2511179

TABLE IV.

Clustering Performance in Terms of NMI, RI, DBI, CONS, and Running Time of Seven Semisupervised Spectral Clustering Algorithms on Synthetic Data Sets

Datasets		Metrics	Algorithms
Datasets		Metrics	SLAM	FCSC	CRM	ASTC	PGCSC	TI	TII
X ₁	5%	NMI-m	0.8908	0.8522	0.8727	0.8876	0.8898	0.8986	0.9016
		NMI-s	0.0016	0.0245	0.0158	0.0016	0.0034	0.0088	0.0024
		RI-m	0.9695	0.9526	0.9533	0.9590	0.9676	0.9714	0.9711
		RI-s	0.0012	0.0121	0.0055	0.0015	0.0007	0.0025	0.0007
		DBI-m	0.6755	0.6917	0.6988	0.6882	0.6783	0.6667	0.5333
		DBI-s	0.0036	0.0110	0.0102	0.0021	0.0030	0.3055	0.3512
		CONS-m	1	0.9870	0.9880	0.9877	1	0.9425	0.9928
		CONS-s	0	0.0225	0.0020	0.0025	0	0.0653	0.0124
		Time-m	1.2133	9.007	0.4925	0.4473	1.3499	0.5223	0.5591
		Time-s	0.0618	0.9381	0.0189	0.1601	0.0523	0.0126	0.0018
	10%	NMI-m	0.8824	0.8845	0.8779	0.8766	0.8829	0.8909	0.8884
		NMI-s	0.0035	0.0054	0.0051	0.0012	0.0051	0.0048	0.0011
		RI-m	0.9660	0.9630	0.9568	0.9560	0.9660	0.9675	0.9675
		RI-s	0.0001	0.0015	0.0007	0.0011	0.0001	0.0025	0.0024
		DBI-m	0.6754	0.6778	0.6841	0.6864	0.6750	0.4215	0.4333
		DBI-s	0.0033	0.0019	0.0010	0.0055	0.0052	0.2646	0.2082
		CONS-m	1	1	0.9823	0.9877	1	0.9747	1
		CONS-s	0	0	0.0083	0.0025	0	0.0438	0
		Time-m	1.1235	9.4850	0.4727	0.3674	1.2161	0.5242	0.5588
		Time-s	0.0446	0.4850	0.0012	0.1587	0.0467	0.0145	0.0104
X ₂	5%	NMI-m	0.9190	0.8544	0.9097	0.9082	0.9156	0.9207	0.9325
		NMI-s	0.0099	0.0836	0.0066	0.0077	0.0066	0.0109	0.0231
		RI-m	0.9758	0.9536	0.9660	0.9654	0.9746	0.9758	0.9794
		RI-s	0.0032	0.0299	0.0036	0.0029	0.0012	0.0032	0.0082
		DBI-m	1.1512	1.6108	1.1583	1.1620	1.1399	1.0206	1.0997
		DBI-s	0.0078	0.8279	0.0082	0.0146	0.0037	0.1732	0.2309
		CONS-m	0.8764	1	0.9890	1	0.8764	1	1
		CONS-s	0.2141	0	0.0010	0	0.2141	0	0
		Time-m	2.7255	17.0034	1.2992	0.5175	2.1397	1.0153	1.0697
		Time-s	0.0878	1.3347	0.6661	0.1098	0.1051	0.0333	0.0268
	10%	NMI-m	0.3648	0.9102	0.9248	0.9253	0.9165	0.9330	0.9366
		NMI-s	0.0673	0.0267	0.0216	0.0128	0.0239	0.0247	0.0266
		RI-m	0.6690	0.9745	0.9705	0.9716	0.9734	0.9785	0.9805
		RI-s	0.0271	0.0080	0.0073	0.0040	0.0084	0.0085	0.0098
		DBI-m	1.0987	1.3430	1.1680	1.1659	1.1362	1.0341	1.0239
		DBI-s	0.1536	0.3446	0.0241	0.0229	0.0302	0.2777	0.1939
		CONS-m	1	1	0.9893	0.9883	1	1	1
		CONS-s	0	0	0.0006	0.0015	0	0	0
		Time-m	2.7928	17.6024	1.1485	0.4837	2.1009	1.0122	1.0683
		Time-s	0.0967	1.4077	0.4569	0.0962	0.0891	0.0293	0.0175
X ₃	5%	NMI-m	0.8956	0.7237	0.8853	0.8868	0.8914	0.9033	0.9035
		NMI-s	0.0111	0.0658	0.0243	0.0166	0.0074	0.0089	0.0091
		RI-m	0.9731	0.9034	0.9621	0.9631	0.9715	0.9753	0.9753
		RI-s	0.0034	0.0340	0.0081	0.0052	0.0025	0.0028	0.0028
		DBI-m	6.4415	8.0839	6.1906	6.4519	6.4266	5.7511	5.7058
		DBI-s	0.1177	3.4620	0.5492	0.3011	0.0402	0.1025	0.1349
		CONS-m	1	1	0.9798	0.9917	1	1	1
		CONS-s	0	0	0.0211	0.0047	0	0	0
		Time-m	3.2666	13.7548	0.7066	0.4121	4.6916	1.6856	1.8101
		Time-s	0.3502	2.2228	0.0481	0.0700	0.0520	0.0886	0.0329
	10%	NMI-m	0.8972	0.7942	0.9011	0.8882	0.9004	0.9015	0.9014
		NMI-s	0.0048	0.1323	0.0052	0.0048	0.0065	0.0027	0.0029
		RI-m	0.9737	0.9255	0.9675	0.9637	0.9742	0.9748	0.9748
		RI-s	0.0016	0.0666	0.0019	0.0016	0.0019	0.0009	0.0009
		DBI-m	6.3587	5.4516	6.0547	6.3895	6.3620	6.2377	6.3255
		DBI-s	0.0868	1.1024	0.2106	0.0746	0.0744	0.2129	0.1745
		CONS-m	1	1	0.9943	1	0.8333	1	1
		CONS-s	0	0	0.0045	0	0.2887	0	0
		Time-m	3.2024	12.3624	0.7028	0.3740	4.2121	1.6924	1.8049
		Time-s	0.0657	1.2427	0.0212	0.2277	0.1241	0.0222	0.0155

Note: *-m and *-s denote the values of the mean and standard deviation, respectively; TI and TII are the separate abbreviations of our proposed TI-APJCSC and TH-APJCSC algorithms.