Research Article

m-Health of Nutrition: Improving Nutrition Services with Smartphone and Machine Learning

Table 4

Performance comparison of six blood pressure prediction models: 26 nutrients intake and 5 personal information.

Prediction modelTrain-test splitSystolic pressure (95% CI)Diastolic pressure (95% CI)
MAERMSEMAERMSE

DT7 : 315.53 (15.46, 15.60)20.60 (20.51, 20.70)10.42 (10.39, 10.46)13.57 (13.51, 13.63)
8 : 215.41 (15.32, 15.50)20.42 (20.30, 20.54)10.40 (10.35, 10.45)13.54 (13.46, 13.62)

KNN7 : 312.16 (12.12, 12.20)16.14 (16.05, 16.23)8.25 (8.23, 8.28)11.01 (10.98, 11.06)
8 : 212.21 (12.17, 12.25)16.18 (16.11, 16.25)8.20 (8.16, 8.24)10.93 (10.86, 10.99)

AdaBoost7 : 313.04 (12.84, 13.24)16.43 (16.23, 16.62)7.95 (7.88, 8.01)10.22 (10.14, 10.29)
8 : 212.84 (12.62, 13.07)16.25 (16.03, 16.46)7.93 (7.86, 8.00)10.19 (10.12, 10.27)

Extra-trees7 : 315.91 (15.39, 16.42)20.46 (19.95, 21.00)10.56 (10.52, 10.60)13.77 (13.70, 13.83)
8 : 215.60 (15.51, 15.68)20.68 (20.57, 20.79)10.49 (10.43, 10.54)13.63 (13.54, 13.72)

DAE + BNN7 : 312.78 (12.68, 12.88)17.39 (17.21, 17.57)7.95 (7.87, 8.04)10.56 (10.44, 10.67)
8 : 212.75 (12.63, 12.86)17.28 (17.11, 17.44)7.84 (7.78, 7.90)10.40 (10.31, 10.48)

GBDT7 : 310.65 (10.62, 10.68)14.22 (14.18, 14.27)7.18 (7.16, 7.19)9.40 (9.38, 9.43)
8 : 210.53 (10.46, 10.60)14.07 (13.97, 14.17)7.15 (7.12, 7.18)9.35 (9.30, 9.40)