Skip to main content

Table 5 The performance of different approaches to the unbalanced dataset

From: A deep learning framework for identifying essential proteins based on multiple biological information

Method

Accuracy

Precision

Recall

F1 score

Specificity

NPV

Sampling

0.9048

0.7306

0.7885

0.7585

0.9320

0.9496

Raw

0.8406

0.5829

0.5442

0.5629

0.9095

0.8956

CW = 2

0.8689

0.7379

0.4735

0.5768

0.9609

0.8870

CW = 3

0.8431

0.5870

0.6864

0.6328

0.8815

0.9197

CW = 4

0.8489

0.6266

0.6083

0.6173

0.9092

0.9026

CW = 5

0.8523

0.5887

0.6964

0.6380

0.8881

0.9271

CW = 6

0.8589

0.6244

0.5991

0.6115

0.9180

0.9096

CW = 7

0.8406

0.5596

0.6920

0.6188

0.8747

0.9251

CW = 8

0.8648

0.6522

0.6904

0.6707

0.9082

0.9217

CW = 9

0.8673

0.5902

0.7094

0.6443

0.8995

0.9382

CW = 10

0.8489

0.6069

0.6709

0.6373

0.8928

0.9167