Interpretable classifiers for prediction of disability trajectories using a nationwide longitudinal database

BMC Geriatrics

Table 3 Performance of machine learning for three-class task prediction

	Full variables					Selected variables with LASSO
	LR	SVM	RF	ANN	XGBoost	LR	SVM	RF	ANN	XGBoost
Accuracy	0.706	0.734	0.773	0.735	0.771	0.714	0.744	0.774	0.744	0.771
Recall	0.779	0.810	0.844	0.785	0.834	0.780	0.810	0.843	0.780	0.840
Precision	0.769	0.807	0.854	0.775	0.833	0.771	0.800	0.848	0.773	0.848
F1 Score	0.760	0.807	0.848	0.766	0.832	0.761	0.802	0.845	0.759	0.843
Hamming	0.221	0.190	0.156	0.215	0.166	0.220	0.190	0.157	0.220	0.160
Jaccard	0.639	0.705	0.759	0.645	0.735	0.639	0.696	0.754	0.636	0.751
Kappa	0.507	0.495	0.568	0.516	0.567	0.512	0.518	0.575	0.514	0.564

Accuracy, recall, precision, F1 score were all calculated with weighted metrics. Hamming, Jaccard, and Kappa refer to Hamming distance, Jaccard similarity coefficient, and Cohen’s kappa score

ISSN: 1471-2318