Table 2.

Summary of best models for each learning strategy, considering the F1 score.

Learning strategy and NLPa feature			Classifier family		AUCb		APc		Precision		Recall		F1		Specificity
Class weighting
	BOWd	SVMe		0.91		0.56		0.50		0.45		0.48		0.98
	EMBf	SVM		0.91		0.61		0.21		0.77		0.33		0.85
Meta-classifier
	BOW	XGBg		0.90		0.45		0.18		0.86		0.30		0.79
	EMB	ADAh		0.92		0.38		0.18		1.00		0.30		0.76
Data augmentation
	EMB	SVM		0.89		0.54		0.35		0.59		0.44		0.94
	EMB	ADA		0.84		0.36		0.43		0.41		0.42		0.97

^aNLP: natural language processing.

^bAUC: area under the curve.

^cAP: average precision.

^dBOW: bag-of-words.

^eSVM: support vector machines.

^fEMB: word embeddings.

^gXGB: extreme gradient boosting.

^hADA: adaptive boosting.