Table 2.

Classification accuracies for the PU dataset with five runs.

	CAP1-PU	CAP2-PU	CNN1-PU	CNN2-PU	RF-PU	SVM-PU
K	0.9324 ± 0.0242	0.9456 ± 0.0181	0.9345 ± 0.0130	0.9332 ± 0.0221	0.6450 ± 0.0147	0.7070 ± 0.0094
OA	0.9490 ± 0.0185	0.9590 ± 0.0138	0.9511 ± 0.0097	0.9496 ± 0.0169	0.7189 ± 0.0124	0.7703 ± 0.0071
AA	0.9542 ± 0.0112	0.9627 ± 0.0138	0.9367 ± 0.0174	0.9563 ± 0.0089	0.7869 ± 0.0098	0.8273 ± 0.0096
C1	0.8943 ± 0.0545	0.9277 ± 0.0264	0.9468 ± 0.0279	0.9382 ± 0.0407	0.6604 ± 0.0143	0.7629 ± 0.0248
C2	0.9686 ± 0.0278	0.9659 ± 0.0210	0.9807 ± 0.0070	0.9507 ± 0.0272	0.6892 ± 0.0237	0.7374 ± 0.0082
C3	0.9229 ± 0.0312	0.8882 ± 0.1190	0.8518 ± 0.0782	0.8851 ± 0.0404	0.6364 ± 0.0284	0.7171 ± 0.0438
C4	0.9797 ± 0.0096	0.9761 ± 0.0121	0.9724 ± 0.0178	0.9720 ± 0.0124	0.9142 ± 0.0250	0.9042 ± 0.0259
C5	1.0000 ± 0.0000	1.0000 ± 0.0000	1.0000 ± 0.0000	1.0000 ± 0.0000	0.9928 ± 0.0052	0.9969 ± 0.0016
C6	0.9609 ± 0.0338	0.9781 ± 0.0284	0.9212 ± 0.0241	0.9483 ± 0.0321	0.7191 ± 0.0334	0.7738 ± 0.0324
C7	0.9831 ± 0.0137	0.9800 ± 0.0085	0.8783 ± 0.1855	0.9620 ± 0.0125	0.8296 ± 0.0205	0.8537 ± 0.0288
C8	0.8796 ± 0.0806	0.9483 ± 0.0224	0.8799 ± 0.0649	0.9504 ± 0.0196	0.6612 ± 0.0332	0.7015 ± 0.0380
C9	0.9985 ± 0.0009	0.9998 ± 0.0005	0.9993 ± 0.0006	1.0000 ± 0.0000	0.9793 ± 0.0072	0.9986 ± 0.0013