MIT Press

Table 4.

Best performance scores for each model + size combination

Model	Size	f_0.5 (macro)	f_0.5 (micro)	f₁ (macro)	f₁ (micro)	f₂ (macro)	f₂ (micro)
DecisionTreeClassifier	s	0.278	0.501	0.302	0.520	0.354	0.542
DecisionTreeClassifier	m	0.248	0.461	0.264	0.483	0.306	0.507
DecisionTreeClassifier	l	0.255	0.477	0.273	0.493	0.317	0.509
ExtraTreesClassifier	s	0.483	0.700	0.341	0.547	0.277	0.449
ExtraTreesClassifier	m	0.453	0.686	0.316	0.520	0.255	0.419
ExtraTreesClassifier	l	0.422	0.673	0.289	0.495	0.232	0.392
RandomForestClassifier	s	0.567	0.738	0.424	0.624	0.352	0.540
RandomForestClassifier	m	0.534	0.728	0.392	0.601	0.323	0.511
RandomForestClassifier	l	0.508	0.719	0.365	0.581	0.299	0.488
LSTMClassifier	s	0.779	0.851	0.739	0.833	0.704	0.816
LSTMClassifier	m	0.756	0.822	0.684	0.796	0.630	0.771
LSTMClassifier	l	0.765	0.831	0.701	0.804	0.651	0.780
MLPClassifier	s	0.780	0.845	0.736	0.833	0.703	0.822
MLPClassifier	m	0.792	0.859	0.753	0.845	0.720	0.832
MLPClassifier	l	0.798	0.861	0.751	0.850	0.721	0.847

Model	Size	f_0.5 (macro)	f_0.5 (micro)	f₁ (macro)	f₁ (micro)	f₂ (macro)	f₂ (micro)
DecisionTreeClassifier	s	0.278	0.501	0.302	0.520	0.354	0.542
DecisionTreeClassifier	m	0.248	0.461	0.264	0.483	0.306	0.507
DecisionTreeClassifier	l	0.255	0.477	0.273	0.493	0.317	0.509
ExtraTreesClassifier	s	0.483	0.700	0.341	0.547	0.277	0.449
ExtraTreesClassifier	m	0.453	0.686	0.316	0.520	0.255	0.419
ExtraTreesClassifier	l	0.422	0.673	0.289	0.495	0.232	0.392
RandomForestClassifier	s	0.567	0.738	0.424	0.624	0.352	0.540
RandomForestClassifier	m	0.534	0.728	0.392	0.601	0.323	0.511
RandomForestClassifier	l	0.508	0.719	0.365	0.581	0.299	0.488
LSTMClassifier	s	0.779	0.851	0.739	0.833	0.704	0.816
LSTMClassifier	m	0.756	0.822	0.684	0.796	0.630	0.771
LSTMClassifier	l	0.765	0.831	0.701	0.804	0.651	0.780
MLPClassifier	s	0.780	0.845	0.736	0.833	0.703	0.822
MLPClassifier	m	0.792	0.859	0.753	0.845	0.720	0.832
MLPClassifier	l	0.798	0.861	0.751	0.850	0.721	0.847