Predictive classification and regression models for bioimpedance vector analysis: Insights from a southern Cuban cohort

Fig. 1:

Fig. 2:

Fig. 3:

Fig. 4:

Fig. 5:

Response models and their respective metrics for the classification of health status and location variables_ The column Class includes Health Status (Cancer, Healthy), Quartile (1, 2, 3, 4), Centile (50, 75, 95, and 100%), and BIVA status (11, 12, 13, 14, 21, 22, 23, 31, 32, 33, 41, 42, 43, 44)_

Response	Model	Accuracy	Class	Precision	Recall	F1-score
Health status	Fine Tree	92.80%	Cancer	0.912	0.951	0.931
			Healthy	0.947	0.905	0.926
BIVA status	Fine Tree	99.50%	11	1.000	1.000	1.000
			12	1.000	1.000	1.000
			13	1.000	1.000	1.000
			14	1.000	1.000	1.000
			21	1.000	1.000	1.000
			22	1.000	1.000	1.000
			23	0.857	0.750	0.800
			31	1.000	1.000	1.000
			32	1.000	1.000	1.000
			33	0.889	0.941	0.914
			41	1.000	1.000	1.000
			42	1.000	1.000	1.000
			43	1.000	1.000	1.000
			44	1.000	1.000	1.000
Quartile	Linear SVM	100%	1	1.000	1.000	1.000
			2	1.000	0.983	0.991
			3	0.990	1.000	0.995
			4	1.000	1.000	1.000
Centile	RUS Boosted Tree	97%	50%	0.988	0.984	0.986
			75%	0.959	0.953	0.956
			95%	0.948	0.938	0.943
			100%	0.957	1.000	0.978

Accuracy parameters of each model_

Response	Model	R2	RMSE	MSE	MAE
Zc (Ω)	Linear	1.000	0.351(Ω2)	0.123 (Ω)	0.233 (Ω)
θc (°)	Linear	0.980	0.237 (°2)	0.056 (°)	0.166 (°)
Xcc (Ω)	Linear SVM	0.990	2.378 (Ω2)	5.652 (Ω)	1.674 (Ω)
Rc (Ω)	Linear SVM	1.000	4.216 (Ω2)	17.773 (Ω)	3.223 (Ω)