Table 4 Overview of input text length and its impact on coding performance with prediction probabilities greater than 0.5.

From: Multi-label text classification via secondary use of large clinical real-world data sets

# of Tokens

> 512

\(<=\) 512

Model

Precision

Recall

F1-measure

MAP

Precision

Recall

F1-measure

MAP

medBERT.de

0.717

0.538

0.579

0.762

0.793

0.767

0.767

0.812

surgeryBERT.at

0.711

0.527

0.572

0.758

0.777

0.743

0.747

0.793

fastText

0.659

0.513

0.544

0.710

0.730

0.719

0.709

0.766

CNN

0.660

0.464

0.511

0.712

0.714

0.674

0.681

0.731

SVM

0.773

0.593

0.639

0.836

0.758

0.725

0.728

0.780

LR

0.557

0.386

0.425

0.593

0.598

0.568

0.571

0.616

Back to article page