Table 1 Distribution of 3-class DILI predictions in cross validations and bootstrapping.

From: Development of Decision Forest Models for Prediction of Drug-Induced Liver Injury in Humans Using A Large Set of FDA-approved Drugs

 

Actual 

Probability of Prediction

No-DILI

Most-DILI

Less-DILI

Unknown

Cross validations

No-DILI

0.6455

0.1029

0.2462

0.0054

Most-DILI

0.2344

0.3793

0.3784

0.0079

Less-DILI

0.2585

0.2083

0.5268

0.0064

Bootstrapping Strategy A

No-DILI

0.6259

0.1136

0.2525

0.0081

Most-DILI

0.2385

0.3660

0.3844

0.0111

Less-DILI

0.2699

0.2120

0.5080

0.0101

Bootstrapping Strategy B

No-DILI

0.6137

0.1176

0.2640

0.0046

Most-DILI

0.2492

0.3532

0.3911

0.0065

Less-DILI

0.2815

0.2117

0.5020

0.0048

Back to article page