Skip to main content

Table 2 Statistical description of 27 datasets.

From: Predictive modeling of structured electronic health records for adverse drug event detection

   

Number of features

Dataset

Instances

% Pos.

Codes

Measurements

Combination

D642

3733

2.87%

3999

494

8262

E273

183

12%

912

240

2935

F110

146

22.6%

1051

205

2958

F112

146

63.7%

1054

205

2963

F130

112

54.5%

779

142

2237

F132

112

27.7%

777

142

2231

F150

111

14.4%

476

107

1543

F151

111

17.1%

475

107

1542

F152

111

69.4%

481

111

1573

F190

168

31.5%

869

160

2454

F192

168

50%

865

160

2447

F199

168

8.93%

866

160

2448

G240

68

20.6%

444

136

1636

G251

194

6.7%

1014

263

3209

G444

908

2.5%

1774

318

4594

G620

382

6%

1624

280

4152

I427

448

5.1%

1341

299

3852

I952

483

8.3%

1654

333

4471

L270

435

35.9%

1297

325

3912

L271

434

11.1%

1286

325

3897

O355

237

35.4%

736

110

1930

T782

1203

8.5%

1625

319

4405

T783

1207

8.6%

1627

319

4408

T784

1213

60.8%

1628

319

4409

T808

391

87.5%

1229

271

3533

T886

715

6.2%

2226

401

5606

T887

716

61.7%

2230

400

5604