A diabetes prediction model based on Boruta feature selection and ensemble learning

BMC Bioinformatics

Table 7 Model evaluation(10CV)

	Accuracy	Recall	F1 Index	Kappa	Precision	MCC
LR	0.719	0.674	0.719	0.357	0.615	0.362
KNN	0.655	0.573	0.655	0.163	0.546	0.184
SVM	0.692	0.640	0.692	0.290	0.579	0.297
NB	0.659	0.614	0.659	0.232	0.524	0.237
DT	0.896	0.884	0.896	0.768	0.871	0.774
RF	0.577	0.505	0.577	0.017	0.386	0.022
PSO-FCM [15]	0.954	0.956	–	–	0.955	0.908
PCA + K-Means + LR [12]	0.973	0.970	0.970	0.942	0.974	0.943
VAE + SAE With CNN [26]	0.923	–	–	–	–	–
K-Means + LR [11]	0.954	0.954	–	0.897	0.954	0.899
Conv-Lstm [24]	0.972	0.939	–	–	–	–
SVC [17]	0.790	0.700	0.715	–	0.731	–
LE [19]	0.750	0.720	0.730	–	0.730	–
X-BLR [29]	0.940	0.940	0.930	–	0.920	–
CGLSTM [30]	0.978	0.896	0.856	–	0.914	–
KFPredict [32]	0.935	0.980	–	–	0.850	–
My model	0.981	0.984	0.980	0.962	0.977	0.965

ISSN: 1471-2105