Weighted k-Prototypes Clustering Algorithm Based on the Hybrid Dissimilarity Coefficient

Table 5

Comparison of cluster accuracy on the Bank dataset.

Parameter settings and sampling rate = 4.8%	k-prototypes	EKACMD	WKPCA
k = 2	0.783195	0.8275	0.88535
k = 4	0.8272	0.838	—
k = 6	0.8205	0.8885	—
k = 8	0.891292	0.904	—
k = 10	0.900089	0.912	—