Skip to main content

Table 5 Comparison between the performance of our HFE method and Fizzy when applied to the CRC datasets, in terms of mean AUC

From: Taxonomy-aware feature engineering for microbiome classification

  

CRC1

#Features

CRC2

#Features

CRC1 + 2

#Features

 

ML

110

97

50

28

100

92

F-JMI

DT

0.555

0.561

0.526

0.594

0.580

0.612

NB

0.670

0.672

0.446

0.350

0.631

0.611

RF

0.737

0.749

0.574

0.418

0.688

0.632

F-MIM

DT

0.535

0.548

0.590

0.557

0.539

0.561

NB

0.632

0.588

0.578

0.422

0.611

0.637

RF

0.618

0.563

0.528

0.374

0.627

0.675

F-mRMR

DT

0.525

0.530

0.458

0.557

0.591

0.623

NB

0.553

0.544

0.615

0.614

0.562

0.559

RF

0.653

0.665

0.593

0.540

0.686

0.705

#Features by NPFS

 

654

167

513

NPFS-MIM

DT

0.634

0.528

0.615

NB

0.628

0.579

0.619

RF

0.746

0.657

0.671

#Features by HFE

 

97

28

92

HFE

DT

0.680

0.660

0.590

NB

0.750

0.583

0.731

RF

0.800

0.980

0.810