Skip to main content
. 2021 Apr 21;6:88. [Version 1] doi: 10.12688/wellcomeopenres.16718.1

Figure 1.

Figure 1.

A) Bootstrap procedure to compare the effect of different features during field selection, n-grams and distributed representations analyses. B) The best-performing features from previous analyses were selected to compare different hyperparameter combinations with 5-fold cross validation. Finally, the best-performing features and hyperparameters were used to apply the pipeline to the final test set.