Representation learning in intraoperative vital signs for heart failure risk prediction

Table 2 Sensitivity (TPR), specificity (TNR), F1 score, accuracy (ACC) of various classifiers

Feature	Methods	TPR	TNR	F1score	ACC
Statistical Feature	Adaboost	0.83	0.83	0.83	0.83
	DT	0.76	0.78	0.77	0.77
	GBDT	0.83	0.85	0.84	0.84
	LR	0.76	0.7	0.73	0.73
	NB	0.72	0.78	0.75	0.76
	RF	0.69	0.96	0.81	0.81
	SVM	0.66	0.96	0.79	0.81
Text Feature	Adaboost	0.68	0.78	0.74	0.74
	DT	0.61	0.78	0.71	0.71
	GBDT	0.68	0.78	0.74	0.74
	LR	0.61	0.93	0.79	0.81
	NB	0.84	0.73	0.78	0.79
	RF	0.65	0.84	0.76	0.76
Image Feature	GRCNN	0.74	0.89	0.83	0.83

The entries in boldface indicate the best results for classifiers in three learning methods. Specifically, these results demonstrate the GBDT classifier achieves the best results in the prediction of heart failure by statistical feature representation. The sensitivity, specificity and accuracy are 83, 85, 84% respectively; the NB classifier achieves the best results in the prediction of heart failure by text feature representation. The sensitivity, specificity and accuracy are 84, 73, 79% respectively; The sensitivity, specificity and accuracy of classification prediction based on convolutional neural network in image feature representation also reaches 89, 78 and 89%, respectively

ISSN: 1472-6947