An ensemble method for estimating the number of clusters in a big data set using multiple random samples

Journal of Big Data

Table 4 Stability validations on the results of the RSPCE algorithm on synthetic datasets with 2 sample sizes, A = 5000 and B = 10,000 (lower value is better)

Dataset	\(n \backslash es\)	APN						ADM
Dataset	\(n \backslash es\)	5%	10%	20%	30%	40%	50%	5%	10%	20%	30%	40%	50%
DS1	A	0.0070	0	0	0	0	0	0.4648	0.4736	0.4666	0.4660	0.4697	0.4603
DS1	B	0.0140	0.0050	0	0	0	0	0.4768	0.4751	0.4640	0.4628	0.4618	0.4599
DS2	A	0.0042	0.0009	0	0	0	0	0.5765	0.5762	0.5802	0.5802	0.5804	0.5800
DS2	B	0.0045	0.0021	0.0011	0.0006	0.0001	0	0.5837	0.5785	0.5810	0.5805	0.5803	0.5801
DS3	A	0.0110	0.0077	0.0044	0.0031	0.0022	0.0013	0.5206	0.5236	0.5274	0.5262	0.5285	0.5304
DS3	B	0.0300	0.0240	0.0215	0.0185	0.0144	0.0086	0.5211	0.5171	0.5208	0.5209	0.5220	0.5253
DS4	A	0.0417	0.0345	0.0270	0.0210	0.0160	0.0123	0.5065	0.5059	0.5089	0.5084	0.5071	0.5073
DS4	B	0.0540	0.0462	0.0370	0.0288	0.0260	0.0240	0.5045	0.5079	0.5082	0.5039	0.5063	0.5027
DS5	A	0.0396	0.0288	0.0234	0.0190	0.0169	0.0124	0.4999	0.5004	0.5015	0.5026	0.5023	0.5022
DS5	B	0.0498	0.0410	0.0306	0.0266	0.0244	0.0226	0.4980	0.4953	0.4977	0.4982	0.4987	0.4994