Research Article

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

Table 5

Performance comparison of six blood pressure prediction models: 26 nutrients intake.

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

DT7 : 319.95 (19.86, 20.05)26.20 (26.08, 26.31)12.06 (12.00, 12.13)15.75 (15.66, 15.83)
8 : 219.89 (19.81, 19.97)26.14 (26.01, 26.27)12.06 (11.96, 12.15)15.70 (15.56, 15.83)

KNN7 : 315.20 (15.15, 15.25)19.85 (19.77, 19.23)9.20 (9.17, 9.22)11.92 (11.87, 11.96)
8 : 215.17 (15.11, 15.24)19.83 (19.75, 19.91)9.18 (9.15, 9.21)11.89 (11.84, 11.94)

AdaBoost7 : 317.15 (16.88, 17.43)20.79 (20.55, 21.03)10.47 (10.16, 10.78)13.01 (12.71, 13.31)
8 : 217.36 (17.11, 17.60)20.98 (20.76, 21.20)11.09 (10.73,11.44)13.58 (13.23, 13.94)

Extra-trees7 : 320.47 (19.54, 21.41)25.82 (25.00, 26.64)12.11 (12.07, 12.15)15.75 (15.68, 15.82)
8 : 219.99 (19.89, 20.08)26.17 (26.08, 26.26)12.07 (12.01, 12.12)15.72 (15.63, 15.81)

DAE + BNN7 : 313.84 (13.73, 13.94)18.81 (18.66, 18.97)8.36 (8.32, 8.41)11.21 (11.12, 11.29)
8 : 213.60 (13.51, 13.69)18.46 (18.36, 18.56)8.33 (8.28, 8.39)11.10 (11.03 11.18)

GBDT7 : 313.73 (13.68, 13.77)18.21 (18.14, 18.28)8.22 (8.19, 8.24)10.94 (10.90, 10.99)
8 : 213.70 (13.64, 13.76)18.18 (18.10, 18.27)8.19 (8.16, 8.22)10.91 (10.85, 10.97)