Machine-learning-based models to predict cardiovascular risk using oculomics and clinic variables in KNHANES

Zhang, Yuqi; Li, Sijin; Wu, Weijie; Zhao, Yanqing; Han, Jintao; Tong, Chao; Luo, Niansang; Zhang, Kun

doi:10.1186/s13040-024-00363-3

BioData Mining

Table 2 The model performance of 25 algorithms on the validation set and external validation set when TyG-index takes three different cut-off values

From: Machine-learning-based models to predict cardiovascular risk using oculomics and clinic variables in KNHANES

	TyG=8						TyG=8.75						TyG=8.93
Model	Acc	AUC	F1	E-AUC	E-Acc	E-F1	Acc	AUC	F1	E-AUC	E-Acc	E-F1	Acc	AUC	F1	E-AUC	E-Acc	E-F1
GradientBoosting	0.789	0.812	0.870	0.809	0.808	0.884	0.820	0.873	0.694	0.863	0.804	0.666	0.874	0.911	0.709	0.901	0.865	0.693
LGBM	0.789	0.809	0.870	0.805	0.806	0.883	0.818	0.871	0.691	0.861	0.804	0.666	0.873	0.910	0.712	0.898	0.864	0.693
AdaBoost	0.785	0.806	0.868	0.803	0.803	0.881	0.816	0.870	0.689	0.859	0.805	0.671	0.872	0.907	0.703	0.896	0.863	0.685
RandomForest	0.784	0.796	0.867	0.791	0.802	0.881	0.813	0.864	0.683	0.854	0.800	0.661	0.870	0.902	0.693	0.891	0.862	0.676
XGBoost	0.781	0.796	0.865	0.793	0.800	0.878	0.812	0.862	0.686	0.848	0.795	0.658	0.867	0.902	0.701	0.888	0.858	0.679
MLP	0.786	0.806	0.868	0.804	0.805	0.882	0.806	0.862	0.679	0.852	0.794	0.654	0.861	0.898	0.684	0.888	0.856	0.671
RidgeCV	0.784	0.797	0.869	0.793	0.798	0.880	0.802	0.856	0.666	0.847	0.792	0.644	0.858	0.891	0.664	0.882	0.851	0.648
RidgeRegression	0.784	0.797	0.869	0.793	0.798	0.880	0.802	0.856	0.664	0.847	0.792	0.641	0.856	0.891	0.655	0.882	0.851	0.642
LogisticRegression	0.786	0.798	0.871	0.796	0.799	0.880	0.801	0.856	0.665	0.847	0.788	0.639	0.857	0.892	0.665	0.882	0.848	0.644
LinearSVM	0.785	0.799	0.870	0.796	0.798	0.879	0.802	0.856	0.666	0.847	0.793	0.645	0.858	0.892	0.669	0.883	0.851	0.652
ExtraTrees	0.783	0.788	0.867	0.787	0.800	0.879	0.807	0.854	0.678	0.846	0.796	0.658	0.863	0.895	0.680	0.885	0.857	0.665
SGD	0.781	0.794	0.869	0.791	0.797	0.880	0.798	0.854	0.644	0.844	0.783	0.617	0.853	0.891	0.655	0.880	0.844	0.636
PassiveAggressive	0.779	0.788	0.872	0.782	0.794	0.882	0.798	0.852	0.646	0.842	0.787	0.625	0.854	0.888	0.648	0.877	0.844	0.624
SVM	0.775	0.765	0.871	0.769	0.794	0.883	0.801	0.847	0.647	0.830	0.784	0.625	0.844	0.875	0.588	0.863	0.837	0.576
Perceptron	0.774	0.786	0.869	0.777	0.790	0.879	0.795	0.847	0.628	0.837	0.785	0.610	0.854	0.884	0.644	0.874	0.844	0.620
Bagging	0.758	0.753	0.845	0.748	0.768	0.853	0.798	0.836	0.656	0.822	0.782	0.633	0.860	0.874	0.678	0.863	0.854	0.669
GaussianNB	0.741	0.767	0.840	0.757	0.737	0.838	0.494	0.813	0.547	0.799	0.498	0.551	0.452	0.843	0.456	0.824	0.447	0.456
QuadraticDiscriminant	0.460	0.704	0.454	0.706	0.447	0.443	0.749	0.804	0.526	0.798	0.747	0.520	0.816	0.835	0.585	0.824	0.815	0.583
LabelPropagation	0.779	0.770	0.870	0.772	0.801	0.885	0.752	0.798	0.514	0.791	0.735	0.495	0.784	0.817	0.300	0.807	0.772	0.267
LabelSpreading	0.776	0.768	0.870	0.770	0.799	0.884	0.748	0.794	0.501	0.787	0.733	0.486	0.778	0.812	0.256	0.801	0.767	0.226
KNeighbors	0.760	0.718	0.849	0.725	0.773	0.860	0.757	0.782	0.585	0.775	0.747	0.576	0.819	0.811	0.563	0.799	0.809	0.541
DecisionTree	0.722	0.614	0.818	0.611	0.722	0.820	0.739	0.712	0.617	0.702	0.731	0.609	0.813	0.759	0.635	0.750	0.807	0.628
BernoulliNB	0.738	0.660	0.838	0.645	0.738	0.839	0.613	0.707	0.567	0.694	0.597	0.571	0.684	0.754	0.537	0.732	0.659	0.521
ExtraTree	0.709	0.597	0.810	0.605	0.722	0.821	0.726	0.693	0.591	0.687	0.719	0.588	0.797	0.730	0.594	0.717	0.785	0.581
Dummy	0.770	0.500	0.870	0.500	0.788	0.882	0.666	0.500	0.000	0.500	0.657	0.000	0.751	0.500	0.000	0.500	0.742	0.000

Back to article page

ISSN: 1756-0381

Contact us

General enquiries: journalsubmissions@springernature.com