Skip to main content

Table 3 Comparison of performance of models trained on various training set sizes using all predictors on the validation set

From: Automated systems to identify relevant documents in product risk management

Training set size

Ratio of useful: non-useful articles

AUC

2

1: 1

0.684

20

1: 9

0.748

36

1: 17

0.699

74

1: 9.57

0.749

112

1: 4.89

0.794