Clustering by Detecting Density Peaks and Assigning Points by Similarity-First Search Based on Weighted K-Nearest Neighbors Graph

Table 4

Comparison of ACC, AMI, and ARI benchmarks for 6 clustering algorithms on real-world datasets.

Algorithm	AMI	ARI	ACC	EC/AC	Par	AMI	ARI	ACC	EC/AC	Par
Iris						Libras movement
DPC-SFSKNN	0.896	0.901	0.962	3/3	6	0.547	0.368	0.510	10/15	8
DPC	0.812	0.827	0.926	3/3	2	0.535	0.304	0.438	9/15	0.5
DBSCAN	0.792	0.754	0.893	—	0.14/9	0.412	0.183	0.385	—	0.96/5
AP	0.764	0.775	0.911	3/3	6	0.364	0.267	0.453	10/15	2.5
FKNN-DPC	0.912	0.922	0.973	3/3	7	0.508	0.308	0.436	10/15	9
K-means	0.683	0.662	0.823	—	3	0.522	0.306	0.449	—	15
	Wine					Parkinsons
DPC-SFSKNN	0.843	0.851	0.951	3/3	6	0.193	0.380	0.827	2/2	6
DPC	0.706	0.672	0.882	3/3	2	0.210	0.114	0.612	2/2	5
DBSCAN	0.612	0.643	0.856	—	0.42/10	0.205	0.213	0.674	—	0.4/6
AP	0.592	0.544	0.781	3/3	6	0.142	0.127	0.669	2/2	15
FKNN-DPC	0.831	0.852	0.949	3/3	7	0.273	0.391	0.851	2/2	5
K-means	0.817	0.838	0.936	—	3	0.201	0.049	0.625	—	2
	WDBC					Ionosphere
DPC-SFSKNN	0.432	0.516	0.857	2/2	6	0.361	0.428	0.786	3/2	7
DPC	0.002	−0.004	0.602	2/2	9	0.238	0.276	0.681	3/2	0.65
DBSCAN	0.397	0.538	0.862	—	0.27/7	0.544	0.683	0.853	—	0.2/7
AP	0.598	0.461	0.854	2/2	40	0.132	0.168	0.706	2/2	15
FKNN-DPC	0.679	0.786	0.944	2/2	7	0.284	0.355	0.752	2/2	8
K-means	0.611	0.730	0.928	—	2	0.129	0.178	0.712	—	2
	Segmentation					Pima-Indians-diabetes
DPC-SFSKNN	0.665	0.562	0.746	6/7	6	0.037	0.083	0.652	2/2	6
DPC	0.650	0.550	0.684	6/7	3	0.033	0.075	0.647	2/2	4
DBSCAN	0.446	0.451	0.550	—	0.25/10	0.028	0.041	0.577	—	0.15/6
AP	0.405	0.436	0.554	7/7	25	0.045	0.089	0.629	3/2	35
FKNN-DPC	0.655	0.555	0.716	7/7	7	0.001	0.011	0.612	2/2	6
K-means	0.583	0.495	0.612	—	6	0.050	0.102	0.668	—	2
	Seeds					Dermatology
DPC-SFSKNN	0.753	0.786	0.919	3/3	7	0.862	0.753	0.808	7/6	6
DPC	0.727	0.760	0.918	3/3	2	0.611	0.514	0.703	4/6	2
DBSCAN	0.640	0.713	0.874	—	0.17/8	0.689	0.690	0.815	—	0.7/3
AP	0.598	0.682	0.896	3/3	10	0.766	0.701	0.762	7/6	5
FKNN-DPC	0.759	0.790	0.924	3/3	8	0.847	0.718	0.768	7/6	7
K-means	0.671	0.705	0.890	—	3	0.796	0.680	0.702	—	6
	Waveform					Waveform (noise)
DPC-SFSKNN	0.355	0.382	0.725	3/3	5	0.267	0.288	0.651	3/3	6
DPC	0.320	0.269	0.586	3/3	0.5	0.104	0.095	0.502	3/3	0.3
DBSCAN	—	—	—	—	—	—	—	—	—	—
AP	—	—	—	—	—	—	—	—	—	—
FKNN-DPC	0.324	0.350	0.703	3/3	5	0.247	0.253	0.648	3/3	5
K-means	0.363	0.254	0.501	-	3	0.364	0.252	0.512	—	3