Skip to main content

Table 2 Result of experiment one: comparison of different models

From: Towards classifying species in systems biology papers using text mining

  

F1

F1-JN

F1-MH

  

P

R

F

P

R

F

P

R

F

 

fly

0.780

0.929

0.848

0.685

0.890

0.777

0.530

0.808

0.640

NB

mouse

0.810

0.680

0.739

0.697

0.620

0.656

0.683

0.410

0.776

 

yeast

0.750

0.727

0.738

0.646

0.515

0.573

0.747

0.657

0.828

 

fly

0.780

0.929

0.848

0.659

0.899

0.761

0.638

0.677

0.657

AdaBoost

mouse

0.810

0.680

0.739

0.697

0.620

0.656

0.696

0.550

0.615

 

yeast

0.750

0.727

0.738

0.649

0.485

0.555

0.640

0.737

0.685

 

fly

0.791

0.919

0.850

0.729

0.869

0.793

0.606

0.838

0.703

Bagging

mouse

0.788

0.670

0.724

0.670

0.670

0.670

0.831

0.490

0.616

 

yeast

0.768

0.768

0.765

0.638

0.515

0.570

0.696

0.717

0.706

 

fly

0.532

0.667

0.592

0.556

0.354

0.432

0.532

0.667

0.592

Decision Table

mouse

0.515

0.520

0.517

0.388

0.870

0.537

0.515

0.520

0.517

 

yeast

0.740

0.545

0.628

0.727

0.081

0.145

0.740

0.545

0.628

 

fly

0.637

0.798

0.709

0.500

0.687

0.579

0.341

0.606

0.436

Decision tree

mouse

0.557

0.640

0.595

0.500

0.550

0.524

0.339

0.400

0.367

 

yeast

0.729

0.434

0.544

0.596

0.313

0.411

1.000

0.040

0.078

 

fly

0.878

0.727

0.796

0.932

0.697

0.798

0.663

0.576

0.603

Logistics Regression

mouse

0.586

0.750

0.658

0.721

0.490

0.583

0.541

0.730

0.621

 

yeast

0.705

0.626

0.663

0.513

0.808

0.627

0.740

0.545

0.628

 

fly

0.762

0.868

0.802

0.688

0.879

0.764

0.503

0.830

0.621

CRF

mouse

0.789

0.700

0.730

0.725

0.640

0.669

0.766

0.390

0.511

 

yeast

0.734

0.725

0.723

0.652

0.566

0.598

0.744

0.650

0.685

 

fly

0.925

0.641

0.757

0.619

0.684

0.650

0.200

0.240

0.217

SVM

mouse

0.403

1.000

0.574

0.309

0.406

0.351

0.250

0.147

0.185

 

yeast

0.636

0.194

0.297

0.400

0.207

0.273

0.356

1.000

0.525

  1. F1: basic features used to train the model were GN, MH, DT, and JN.
  2. F1-JN: the features used were GN, MH and DT.
  3. F1-MH: the features used were GN, DT, and JN.