Table 1 Model performance based on 10-fold cross-validation.

	Accuracy	Precision	Recall	F1	AUC
Full feature set
GB	0.875 ± 0.025	0.895 ± 0.033	0.865 ± 0.031	0.872 ± 0.025	0.938 ± 0.015
GNB	0.851 ± 0.025	0.821 ± 0.032	0.924 ± 0.021	0.863 ± 0.021	0.862 ± 0.023
LR	0.859 ± 0.021	0.870 ± 0.025	0.859 ± 0.023	0.857 ± 0.021	0.921 ± 0.015
RF	0.870 ± 0.021	0.870 ± 0.026	0.886 ± 0.032	0.871 ± 0.022	0.937 ± 0.014
SVM	0.856 ± 0.022	0.876 ± 0.028	0.845 ± 0.027	0.852 ± 0.023	0.921 ± 0.017
Reduced feature set
GB	0.799 ± 0.029	0.800 ± 0.035	0.819 ± 0.032	0.801 ± 0.029	0.860 ± 0.031
GNB	0.781 ± 0.022	0.765 ± 0.028	0.840 ± 0.043	0.792 ± 0.024	0.800 ± 0.021
LR	0.795 ± 0.030	0.797 ± 0.035	0.814 ± 0.030	0.797 ± 0.029	0.857 ± 0.032
RF	0.805 ± 0.024	0.802 ± 0.033	0.833 ± 0.023	0.809 ± 0.022	0.867 ± 0.025
SVM	0.812 ± 0.027	0.822 ± 0.036	0.816 ± 0.032	0.811 ± 0.027	0.864 ± 0.026

LR Logistic Regression, GNB Gaussian Naïve Bayes, SVM Support Vector Machine, RF Random Forest, GB Gradient Boosting.

Quick links

Search