Table 10.

Best macro-avg results for each prediction schema in the Multi-Class scenario.

Prediction Schema	Feature(s)	Classifier(s)	Resampling(s)	F1-Score
Individual	LBP	MLP	RENN or AllKNN	0.6491
Early Fusion	BSIF & EQP & LPQ	MLP	TomekLink	0.5563
Late Fusion (Top-5)	LBP	MLP	AllKNN & RENN	0.6491
Late Fusion (Top-Features)	BSIF & LBP	MLP	RENN & SMOTE-B2	0.6491
Late Fusion (Top-Classifiers)	LDN & LETRIST	DT & KNN-3	RENN or None	0.4500