Table 6 Performance of manual Gene Ontology rules on the CRAFT corpus

Cellular Component (CC)

Baseline (B1)

X

X

5,532

452

2822

0.925

0.662

0.772

Baseline (B2)

X

X

5,532

452

2822

0.925

0.662

0.772

Syntactic recursion rules

23

21

5,532

452

2,822

0.925

0.662

0.772

Both rules

4,083

724

6,585

969

1,769

0.872

0.788

0.828

Molecular Function (MF)

Baseline (B1)

X

X

337

146

3,843

0.698

0.081

0.145

Baseline (B2)

X

X

1,772

964

2,408

0.648

0.424

0.512

Syntactic recursion rules

11,637

7,353

1,759

977

2,421

0.643

0.421

0.509

Both rules

14,413

7,401

2,422

1,074

1,758

0.693

0.579

0.631

Biological Process (BP)

Baseline (B1)

X

X

4,909

5,682

12,004

0.464

0.290

0.357

Baseline (B2)

X

X

4,913

5,951

12,000

0.452

0.291

0.354

Syntactic recursion rules

182,617

6,847

5,120

6,158

11,793

0.454

0.303

0.363

Both rules

272,535

8,675

9,604

8,464

7,309

0.532

0.568

0.549

All Gene Ontology

Baseline (B1)

X

X

10,778

6,280

18,669

0.632

0.366

0.464

Baseline (B2)

X

X

12,217

7,367

17,230

0.624

0.415

0.498

Syntactic recursion rules

194,277

14,221

12,411

7,588

17,036

0.621

0.422

0.502

Both rules

291,031

16,800

18,611

10,507

10,836

0.640

0.632

0.636