Unlocking the link: predicting cardiovascular disease risk with a focus on airflow obstruction using machine learning

Table 4 Comparisons of six machine learning classifiers

	Accuracy		Precision		Recall		F1 Score		Brier Score
	Train set	Test set	Train set	Test set	Train set	Test set	Train set	Test set	Train set	Test set
Full Population CVD Diagnostic Model
XGBoost	0.7157	0.6665	0.6969	0.2524	0.7634	0.7062	0.7286	0.3719	0.1911	0.2012
Random Forest	0.6928	0.6433	0.6709	0.2408	0.7570	0.7211	0.7113	0.3611	0.2067	0.2110
KNN	1.0000	0.7350	1.0000	0.2245	1.0000	0.3650	1.0000	0.2780	0. 0000	0.2172
Naive Bayes	0.6709	0.6591	0.6650	0.2508	0.6888	0.7240	0.6767	0.3725	0. 2118	0.2107
Catboost	0.7581	0.6765	0.7314	0.2525	0.8159	0.6706	0.7713	0.3669	0.1720	0.1950
LightGBM	0.7166	0.6653	0.6967	0.2495	0.7674	0.6944	0.7303	0.3671	0.1883	0.2013
Airflow Obstruction Population CVD Diagnostic Model
XGBoost	0.7671	0.6539	0.7521	0.3085	0.7968	0.5321	0.7738	0.3906	0.1608	0.2162
Random Forest	0.9831	0.6902	0.9672	0.2977	1.000	0.3578	0.9833	0.3250	0.0189	0.1983
MLP	0.9468	0.6539	0.9367	0.2750	0.9584	0.4037	0.9474	0.3271	0.0417	0.2930
KNN	1.0000	0.6252	1.0000	0.2570	1.0000	0.4220	1.0000	0.3194	0.0000	0.2840
Naive Bayes	0.6786	0.5698	0.6368	0.2929	0.8313	0.7523	0.7211	0.4216	0.2106	0.2643
LightGBM	0.7748	0.6444	0.7582	0.2941	0.8069	0.5046	0.7818	0.3716	0.1587	0.2189

Abbreviations: XGBoost, eXtreme Gradient Boosting; LightGBM, Light Gradient Boosting Machine; KNN, K-Nearest Neighbor; MLP, Multi-Layer Perceptron; CatBoost, Categorical Boosting

ISSN: 1472-6947