Comparing human milk macronutrients measured using analyzers based on mid-infrared spectroscopy and ultrasound and the application of machine learning in data fitting

Table 3 The training data and the test data of the adjusted ultrasound results and the model fitting parameters

	Data set (n = 465)	Adjusted ultrasound	95% LOA^a	Mean squared error (MSE)	Variance score	Lin’s Concordance Correlation Coefficient	P-value^b	Model MSE	Model Bias	Model Variance
Fat (g/dl)	Training data (n = 372)	3.14 ± 0.49	-0.91 ~ 0.91	0.217	0.551	0.67	0.13	0.233	0.231	0.002
Fat (g/dl)	Test data (n = 93)	3.23 ± 0.57	-0.93 ~ 0.97	0.231	0.670	0.67	0.13	0.233	0.231	0.002
Energy (kj/dl)	Training data (n = 372)	270.26 ± 20.71	-41.42 ~ 41.42	445.401	0.499	0.71	0.60	333.67	330.35	3.321
Energy (kj/dl)	Test data (n = 93)	271.52 ± 19.20	-34.92 ~ 36.58	329.751	0.465	0.78	0.60	333.67	330.35	3.321

^a refers to the result of Bland-Altman analysis between adjusted ultrasonic and MIR values; ^b refers to the p-value of t-test between training data and test data

ISSN: 1471-2393