Table 2 The benchmark results of independent test on DS1_Indp and DS2_Indp.

From: AAGP integrates physicochemical and compositional features for machine learning-based prediction of anti-aging peptides

 

Model

Accuracy

Precision

Recall

Specificity

F1-score

AUC

MCC*

DS1_Indp

LGBM

0.955

0.885

0.575

0.993

0.697

0.963

0.692

CB

0.943

0.660

0.775

0.960

0.713

0.963

0.684

GBC

0.952

0.880

0.550

0.993

0.677

0.949

0.674

ET

0.952

0.880

0.550

0.993

0.677

0.965

0.674

RF

0.943

0.683

0.700

0.968

0.691

0.959

0.660

LDA

0.932

0.625

0.625

0.963

0.625

0.914

0.588

KNN

0.936

0.875

0.350

0.995

0.500

0.874

0.530

QDA

0.893

0.446

0.725

0.910

0.552

0.903

0.514

MLP

0.923

0.800

0.200

0.995

0.320

0.860

0.376

DS2_Indp

ET

0.941

0.818

0.450

0.990

0.581

0.808

0.580

CB

0.939

0.933

0.350

0.998

0.509

0.814

0.551

RF

0.936

0.833

0.375

0.993

0.517

0.829

0.533

LGBM

0.936

0.929

0.325

0.998

0.481

0.788

0.528

GBC

0.932

0.750

0.375

0.988

0.500

0.808

0.500

LDA

0.930

0.800

0.300

0.993

0.436

0.704

0.463

MLP

0.925

0.652

0.375

0.980

0.476

0.748

0.459

KNN

0.916

0.560

0.350

0.973

0.431

0.713

0.400

QDA

0.898

0.439

0.450

0.943

0.444

0.720

0.388

  1. *The models are ranked according to MCC.