. 2019 Mar 21;9(1):33. doi: 10.3390/diagnostics9010033

Table 3.

Absolute and relative errors of all algorithms tested.

Absolute and Relative Errors of Predicted Lung Function Values
		Mean Prediction		Median Prediction		kNN		Polynomial Regression		Neural Network		XGBoost
		MAR	MRE	MAR	MRE	MAR	MRE	MAR	MRE	MAR	MRE	MAR	MRE
inspiration	%FEV₁	21 ± 13	46 ± 43	20 ± 16	39 ± 6	19 ± 13	43 ± 39	19 ± 13	40 ± 37	20 ± 13	41 ± 34	22 ± 18	47 ± 50
	%FEV₁/VC	12 ± 8	22 ± 17	11 ± 8	21 ± 12	9 ± 8	17 ± 15	10 ± 7	19 ± 15	12 ± 8	22 ± 17	10 ± 8	19 ± 16
	RV/TLC	14 ± 10	30 ± 40	14 ± 10	31 ± 42	13 ± 9	28 ± 38	12 ± 8	26 ± 36	13 ± 9	27 ± 39	14 ± 12	31 ± 46
expiration	%FEV₁	21 ± 13	46 ± 43	20 ± 16	39 ± 6	18 ± 15	36 ± 37	17 ± 13	36 ± 37	19 ± 14	41 ± 36	22 ± 16	46 ± 45
	%FEV₁/VC	12 ± 8	22 ± 17	11 ± 8	21 ± 12	9 ± 8	16 ± 14	10 ± 7	18 ± 15	12 ± 8	22 ± 17	10 ± 8	20 ± 17
	RV/TLC	14 ± 10	30 ± 40	14 ± 10	31 ± 42	11 ± 9	25 ± 31	11 ± 8	22 ± 28	11 ± 8	22 ± 25	12 ± 9	24 ± 26
delta	%FEV₁	21 ± 13	46 ± 43	20 ± 16	39 ± 6	18 ± 15	39 ± 42	19 ± 15	40 ± 37	21 ± 19	38 ± 35	18 ± 15	40 ± 46
	%FEV₁/VC	12 ± 8	22 ± 17	11 ± 8	21 ± 12	9 ± 8	16 ± 14	9 ± 8	17 ± 15	20 ± 16	35 ± 25	9 ± 8	17 ± 16
	RV/TLC	14 ± 10	30 ± 40	14 ± 10	31 ± 42	12 ± 9	25 ± 30	12 ± 10	25 ± 30	26 ± 22	47 ± 49	14 ± 12	29 ± 34

Mean absolute error (MAR) and the mean relative error (MRE) of the predicted lung function values. The MAR is given as the amount of the absolute error needed to compensate the negative signs. FEV1: forced expiratory volume in one second; VC: vital capacity; RV: residual volume; TLC: total lung capacity; kNN: k-nearest neighbour; MRE: mean relative error; MAR: mean absolute error.