Cough recognition in pneumoconiosis patients based on a flexible patch with an embedded ACC sensor for remote monitoring

Table 2 The classification results with different standard classification algorithms

Performance evaluation		RandomForest	Extra Trees		CatBoost	Light GBM	Gradient Boosting	KNN	XGBoost
Accuracy	P	0.943	0.950	0.943		0.940	0.937	0.938	0.938
	H	0.886	0.891	0.894		0.889	0.891	0.897	0.880
	C	0.923	0.924	0.925		0.921	0.917	0.918	0.922
Precision	P	0.953	0.953	0.946		0.943	0.940	0.942	0.941
	H	0.901	0.905	0.901		0.894	0.900	0.904	0.887
	C	0.919	0.933	0.927		0.922	0.917	0.921	0.924
Recall	P	0.950	0.948	0.943		0.940	0.938	0.938	0.938
	H	0.889	0.881	0.892		0.887	0.886	0.895	0.879
	C	0.927	0.925	0.925		0.921	0.913	0.918	0.922
F1 score	P	0.948	0.950	0.947		0.943	0.942	0.942	0.942
	H	0.877	0.883	0.883		0.879	0.877	0.889	0.869
	C	0.927	0.920	0.925		0.922	0.911	0.918	0.922
AUC	P	0.986	0.985	0.987		0.984	0.985	0.976	0.985
	H	0.949	0.960	0.959		0.955	0.951	0.959	0.947
	C	0.978	0.979	0.977		0.978	0.975	0.969	0.975
Performance evaluation		Bagging	SVM		AdaBoost	Logistic Regression	Decision Tree		Naive Bayes
Accuracy	P	0.934	0.927	0.922		0.934	0.906		0.815
	H	0.875	0.866	0.867		0.864	0.815		0.853
	C	0.912	0.908	0.905		0.906	0.878		0.813
Precision	P	0.941	0.929	0.926		0.936	0.910		0.843
	H	0.898	0.895	0.926		0.882	0.831		0.873
	C	0.918	0.912	0.906		0.909	0.882		0.832
Recall	P	0.929	0.927	0.923		0.934	0.907		0.821
	H	0.875	0.862	0.863		0.860	0.814		0.849
	C	0.913	0.908	0.905		0.906	0.874		0.814
F1 score	P	0.945	0.929	0.925		0.936	0.912		0.791
	H	0.854	0.837	0.855		0.840	0.817		0.826
	C	0.903	0.904	0.905		0.903	0.880		0.788
AUC	P	0.972	0.980	0.973		0.981	0.902		0.928
	H	0.920	0.890	0.936		0.893	0.830		0.839
	C	0.966	0.948	0.968		0.955	0.878		0.913

(P: Patients; H: Healthy adults; C: the combined group of healthy adults and patients)

ISSN: 1472-6947