COIN: Correlation Index-Based Similarity Measure for Clustering Categorical Data

Table 7

Comparative analysis of clustering accuracy of COIN algorithm with other five existing algorithms on selected nine datasets.

S. no.	Datasets/Algorithms	MMR	K-ANMI	G-ANMI	COOLCAT	MGR	COIN	Avg CPUTIME in seconds
1	Zoo	0.911	0.733	0.874	0.785	0.931	0.920	0.2
2	Hayes-Roth	0.386	0.447	0.409	0.470	0.485	0.553	0.12
3	Votes	0.687	0.869	0.871	0.839	0.828	0.887	0.74
4	Balance Scale	0.635	0.506	0.518	0.506	0.635	0.755	2.21
5	Breast Cancer	0.669	0.978	0.966	0.650	0.884	0.771	2.81
6	Car Evaluation	0.7	0.7	0.7	0.7	0.7	0.736	13.92
7	Chess	0.523	0.547	0.543	0.533	0.534	0.643	80.00
8	Mushroom	0.84	0.587	0.547	0.531	0.677	0.827	506
9	Nursery	0.441	0.61	0.372	0.569	0.53	0.657	404