Skip to main content

Table 2 Macro-averaging F1-score/precision/recall of four typical classifiers based on negative samples selected by the proposed negative sample selection method and randomly selected negative samples

From: Inverse similarity and reliable negative samples for drug side-effect prediction

Method

Classifier

Macro_ F1

Macro_P

Macro_R

ComNegative

KNN

0.800

0.728

0.772

ComRandom

KNN

0.619

0.598

0.651

ChemTarRandom

KNN

0.581

0.553

0.626

ComNegative

ELM

0.774

0.761

0.604

ComRandom

ELM

0.586

0.572

0.601

ChemTarRandom

ELM

0.572

0.561

0.585

ComNegative

SVM

0.898

0.938

0.861

ComRandom

SVM

0.655

0.670

0.642

ChemTarRandom

SVM

0.606

0.622

0.598

ComNegative

RBF

0.822

0.828

0.818

ComRandom

RBF

0.606

0.592

0.622

ChemTarRandom

RBF

0.571

0.561

0.583

  1. The best performance for each classifier is showed in boldface