Machine learning based model for the early detection of Gestational Diabetes Mellitus

Table 7 Most Important 26 Biomarkers Identified by the Model

Feature	Case-Mean	Case-STDdev	Control-Mean	Control-STDdev	p-value
History of high glucose level/diabetes	0.508	0.504	0.000	0.000	0.0000
HbA1C %	5.356	0.498	5.005	0.261	0.0000
Glucose	4.790	1.062	4.271	0.532	0.0001
MCH	27.517	2.428	28.729	2.060	0.0009
NT pro-BNP	46.681	33.107	60.274	36.233	0.0120
HOMA-IR (22.5 Scale)	6.486	11.355	3.233	5.439	0.0148
HOMA-IR (405 Scale)	0.360	0.631	0.180	0.302	0.0148
Insulin	25.421	32.277	16.047	23.941	0.0263
Magnesium	0.726	0.057	0.745	0.060	0.0299
Uric Acid	189.730	40.612	176.887	43.085	0.0378
C-Peptide	3.035	2.037	2.436	1.954	0.0403
Triglyceride	1.746	0.600	1.576	0.663	0.0598
Urea	2.337	0.569	2.489	0.627	0.0697
Chloride	101.127	1.680	101.653	2.413	0.0734
Fibrinogen	4.384	0.584	4.218	0.805	0.0880
MCHC	33.716	1.140	33.971	1.081	0.0904
ALT	11.865	6.613	13.823	10.667	0.1038
Family History of Diabetes	1.000	0.000	0.987	0.115	0.1807
Vit B12	242.581	110.993	230.460	96.509	0.2469
TSH	1.955	1.170	2.088	1.208	0.2577
Potassium	4.046	0.236	4.065	0.225	0.3122
Alk Phos	59.524	16.419	58.173	18.640	0.3276
Cholesterol	5.410	1.025	5.355	1.216	0.3900
FT4	13.798	2.722	13.753	1.549	0.4509
Homocysteine Plasma LC-MSMS	5.919	1.763	5.951	1.258	0.4513
Monocyte Auto	0.507	0.162	0.506	0.137	0.480

ISSN: 1472-6947