-Nearest Neighbor Intervals Based AP Clustering Algorithm for Large Incomplete Data

Table 1

Averaged results of 30 trials using incomplete Iris dataset.

Missing rate (%)	Misclassification ratio (%)					Fowlkes-Mallows index
	IPDS	1NN	KNNM	KNNI	KNNI-FCM	IPDS	1NN	KNNM	KNNI	KNNI-FCM
0	10	7.33	7.33	7.33	10.67	0.8306	0.8668	0.8668	0.8668	0.8196
5	10.37	8.51	8.71	7.69	10.53	0.8259	0.8504	0.8477	0.8616	0.8214
10	8.96	8.40	8.67	8.13	10.56	0.8461	0.8516	0.8479	0.8553	0.8209
15	9.98	8.76	8.87	8.49	10.56	0.8342	0.8469	0.8454	0.8504	0.8227
20	10.51	9.24	9.31	8.44	10.27	0.8277	0.8397	0.8392	0.8512	0.8262
25	9.16	9.40	9.40	9.02	10.62	0.8443	0.8377	0.8375	0.8428	0.8235