Predicting outcomes in older ED patients with influenza in real time using a big data-driven and machine learning approach to the hospital information system

Table 2 Comparisons of predictive accuracies among random forest, logistic regression, KNN, SVM, LightGBM, MLP, and XGBoost in the outcomes of testing dataset of older ED patients with influenza

Outcomes and predictive models	Accuracy	Sensitivity	Specificity	PPV	NPV	AUC
Hospitalization
Random forest	0.769	0.744	0.791	0.762	0.775	0.840
Logistic regression	0.737	0.751	0.726	0.711	0.764	0.799
KNN	0.736	0.737	0.736	0.715	0.757	0.790
SVM	0.750	0.751	0.749	0.728	0.770	0.840
LightGBM	0.748	0.714	0.780	0.744	0.752	0.823
MLP	0.733	0.702	0.760	0.724	0.740	0.806
XGBoost	0.721	0.705	0.736	0.706	0.735	0.800
Pneumonia
Random forest	0.679	0.681	0.679	0.562	0.778	0.765
Logistic regression	0.662	0.661	0.662	0.542	0.764	0.709
KNN	0.645	0.700	0.613	0.522	0.771	0.683
SVM	0.657	0.700	0.631	0.534	0.777	0.733
LightGBM	0.653	0.700	0.625	0.530	0.775	0.724
MLP	0.660	0.660	0.660	0.540	0.762	0.688
XGBoost	0.674	0.700	0.658	0.553	0.784	0.744
Sepsis or septic shock
Random forest	0.795	0.750	0.798	0.179	0.982	0.857
Logistic regression	0.799	0.750	0.801	0.182	0.982	0.832
KNN	0.714	0.750	0.712	0.133	0.980	0.785
SVM	0.707	0.750	0.705	0.130	0.980	0.806
LightGBM	0.739	0.739	0.739	0.143	0.980	0.822
MLP	0.730	0.728	0.730	0.137	0.979	0.761
XGBoost	0.744	0.739	0.744	0.146	0.980	0.811
ICU admission
Random forest	0.860	0.722	0.862	0.054	0.996	0.885
Logistic regression	0.720	0.778	0.719	0.030	0.997	0.867
KNN	0.607	0.611	0.607	0.017	0.993	0.622
SVM	0.768	0.778	0.768	0.036	0.997	0.778
LightGBM	0.809	0.722	0.810	0.040	0.996	0.874
MLP	0.629	0.611	0.629	0.018	0.993	0.649
XGBoost	0.912	0.722	0.914	0.085	0.997	0.902
In-hospital mortality
Random forest	0.792	0.806	0.792	0.079	0.995	0.875
Logistic regression	0.816	0.806	0.816	0.089	0.995	0.889
KNN	0.652	0.639	0.652	0.039	0.988	0.663
SVM	0.789	0.722	0.791	0.071	0.992	0.762
LightGBM	0.769	0.722	0.770	0.065	0.992	0.844
MLP	0.675	0.667	0.675	0.044	0.989	0.728
XGBoost	0.751	0.806	0.750	0.067	0.994	0.858

KNN K-nearest neighbors; SVM Support vector machine; LightGBM Light gradient boosting machine; MLP Multilayer perceptron, XGBoost Extreme Gradient Boosting; ED Emergency department; PPV Positive predictive value; NPV Negative predictive value; AUC Area under the curve

ISSN: 1471-2318