Table 1 Stepwise binary logistic regression results for the 26 weighted-importance STW features.

From: Machine learning classification of early-stage Parkinson’s disease using sit-to-walk biomechanical features

STEP

STW features (cut-off value)

B(SE)

OR (95% CI)

p value

RN2

1

TP_COM_Speed (0.24 m/s)

3.209 (0.690)

24.756 (6.401–95.752)

< 0.001

0.592

2

TP_COM_Speed (0.24 m/s)

3.229 (0.749)

25.261 (5.816–109.712)

< 0.001

0.679

P2_AP_COP-COM (31.97 cm)

1.142 (0.365)

3.133 (1.533–6.400)

0.002

3

TP_COM_Speed (0.24 m/s)

3.923 (0.940)

50.562 (8.008–319.257)

< 0.001

0.718

P2_AP_COP-COM (31.97 cm)

1.227 (0.414)

3.412 (1.515–7.684)

0.003

P2_T10_For_ROM (17.78°)

− 0.996 (0.431)

0.369 (0.159–0.860)

0.021

  1. This model adjusted for age, sex, height, and body mass index; STW: Sit-to-walk; B: Logistic regression coefficient; SE: Standard error; OR: Odds ratio; CI: Confidence interval; RN2 is the fit statistic for the Nagelkerke model; TP: Total phase; COM: Center of mass; P2: Phase 2; AP: Anteroposterior; COP-COM: Distance between center of press and center of mass; T10: the 10th thoracic vertebra; For: Forward; ROM: Range of motion; Boldface indicates significant differences; p < 0.05.
Back to article page