Skip to main content
. 2021 Apr 12;9:e11199. doi: 10.7717/peerj.11199

Table 3. Linear regression (LR) coefficients following cross validation on the training subset.

Peak vGRF [BW] y= 2.23 + 0.15*Speed + 0.33* Accel-based Est. −0.34*Step Freq.
B SE t p
Intercept 2.23 0.46 4.87 0.000
Speed [m/s] 0.15 0.03 4.42 0.000
Acceleration-based Estimate [BW] 0.33 0.05 6.54 0.000
Step Frequency [Hz] −0.34 0.12 −2.84 0.006
Body Mass [kg] 0.001 0.003 0.50 0.621
Vertical Impulse [BW*s] y= 0.69 − 0.10*Step Freq.
B SE t p
Intercept 0.69 0.02 31.13 0.000
Speed [m/s] −0.002 0.001 −1.73 0.089
Acceleration-based Estimate [BW*s] −0.05 0.03 −1.79 0.079
Step Frequency [Hz] −0.10 0.004 −25.95 0.000
Body Mass [kg] 0.00003 0.00009 0.32 0.752
Contact Time [s] y= 0.230 − 0.019*Speed + 0.151*Accel-based Est. + 0.0007*Body Mass
B SE t p
Intercept 0.230 0.033 6.88 0.000
Speed [m/s] −0.019 0.002 −8.62 0.000
Acceleration-based Estimate [s] 0.151 0.063 2.42 0.019
Step Frequency [Hz] −0.011 0.007 −1.51 0.135
Body Mass [kg] 0.0007 0.0002 4.05 0.000

Notes.

Unstandardized coefficients (B), coefficient standard errors (SE), t values (t), and p values (p) are listed for independent variables used to predict peak vertical ground reaction force (vGRF), vertical impulse, and contact time, where Accel-based Est. is the Acceleration-based estimate and Freq. is Frequency. Equations with statistically significant predictor variables are included. Bold p values indicate p < 0.05.