Density Peaks Clustering Based on Feature Reduction and Quasi-Monte Carlo

Table 4

The comparison of all algorithms in terms of SC and CH on unlabeled data sets.

Algorithm	Flame		Aggregation		S2		KDD
	SC	CH	SC	CH	SC	CH	SC	CH
QMC-DPC (PCA)	0.379	153.672	0.461	1032.876	0.452	15 011.749	0.941 5	37 569.031 7
QMC-DPC (AE)	0.303	129.028	0.493	1202.934	0.710	19 744.353	0.944 9	45 484.132 6
QMC-DPC (t-SNE)	0.409	190.938	0.493	1202.934	0.704	19 149.117	—	—
DPC-KNN-PCA	0.403	178.217	0.281	329.073	0.710	19 744.353	—	—
DLORE-DP	0.323	108.244	0.463	883.713	0.611	17 689.904	—	—
SNN-DPC	0.132	37.009	0.363	851.953	0.002	1124.993	—	—
DPC	0.347	133.615	0.493	1202.934	0.710	19 744.353	—	—
DBSCAN	0.295	6.894	0.318	606.285	0.529	3178.635	0.229	18 701.700
AP	0.334	247.017	0.421	660.422	0.496	17 747.519	—	—
K-means	0.245	80.406	0.425	967.394	0.539	7293.533	0.694	13 042.686

The bold values mean the best results.