VIASCKDE Index: A Novel Internal Cluster Validity Index for Arbitrary-Shaped Clusters Based on the Kernel Density Estimation

Table 8

The best parameters for the datasets that were detected by the cluster validity indices with the Spectral Clustering algorithm are given in Table 7.

Dataset	Spectral clustering parameters	Best parameters detected by indices for the Spectral Clustering algorithm
		SI	DI	DB	CH	S_Dbw	DSI	RMSSTD	VIASCKDE
Half-kernel	n_clusters	14	2	15	15	14	15	2	2
Two spirals	n_clusters	15	2	15	15	15	15	2	2
Outlier	n_clusters	2	4	4	13	3	4	2	4
Corners	n_clusters	12	4	12	12	15	14	2	2
Cluster in cluster	n_clusters	4	2	4	15	15	15	2	2
Crescent full moon	n_clusters	5	2	5	13	15	14	2	6
Moon	n_clusters	15	2	15	15	15	15	2	2
Face	n_clusters	11	2	10	12	15	13	2	2
Wave	n_clusters	7	2	15	15	15	15	2	2
Fisher iris	n_clusters	2	2	2	3	15	2	2	3
Breast cancer	n_clusters	2	2	2	2	11	14	15	12
Aggregation	n_clusters	4	2	6	14	2	15	2	2
Thyroid	n_clusters	3	2	3	3	15	15	2	3
Zelnik1	n_clusters	12	2	13	12	15	13	3	3
Zelnik5	n_clusters	8	2	8	15	15	15	2	4
Xclara	n_clusters	3	2	3	3	10	3	2	3
Banana	n_clusters	9	2	9	15	14	15	2	2
Ds2c2sc13	n_clusters	3	3	5	8	2	15	2	5
2sp2glob	n_clusters	7	2	15	15	15	15	2	7
Cure-t1-2000n	n_clusters	5	2	4	13	2	12	2	3