Table 20.

Performance comparison of our method vs. mRMR for the imbalanced E. coli data set when identical numbers of features are selected.

	AUC		Precision		Recall		F1		MCC
N4	0.691	0.651	0.725	0.678	0.280	0.269	0.404	0.385	0.391	0.363
N5	0.690	0.675	0.737	0.687	0.295	0.254	0.421	0.371	0.407	0.356
N6	0.701	0.681	0.742	0.708	0.287	0.220	0.414	0.336	0.403	0.338
N7	0.714	0.686	0.735	0.712	0.275	0.212	0.400	0.326	0.392	0.333
N8	0.705	0.692	0.742	0.713	0.288	0.209	0.415	0.323	0.405	0.330
N9	0.707	0.692	0.726	0.713	0.293 >	0.199	0.417 >	0.312	0.401	0.322
N10	0.711	0.697	0.724	0.703	0.294 >	0.193	0.418 >	0.302	0.401	0.313
N11	0.714 <	0.702	0.732	0.697	0.278 >	0.187	0.403	0.295	0.393	0.306
N12	0.712 <	0.704	0.725	0.683	0.292	0.192	0.416	0.300	0.400	0.305
N13	0.714 <	0.715	0.733	0.713	0.287	0.250	0.413	0.370	0.400	0.360