We built a first model including partner's BMI as a response variable and with a unique covariate corresponding to direct preferences for BMI.
For instance, we built a second model for Vm = −70 mV, where we assumed a recoil to an intermediate state that increases Esp by an extra 1.4 kcal/mol, as opposed to retracting to the original stretching of the depolarized state with Esp = 0.
We built a second model adding variables from the data mining procedure.
As education level was missing in 18% of cases and is difficult to assess in women <25 years, we built a fourth model with education restricted to women aged ≥25 years.
In MEBC, we simply assign more weights on the most probable classes for correcting biases [15, 34], while in RFBC, we build a second RF model to estimate the biases generated from the original RF model.
To prove its worth, Reaction Engines plans to build a fifth-scale-model and fly it.
This set was used to build a second PLS-DA model.
Only those variables with RSDs <50% were included in a truncated data set used to build a second PLS-DA model.
Other techniques to investigate the presence of anisotropies in a material are present in literature (see e.g. [15]), that we don't apply here, since we are investigating only the presence of main sources of anisotropy, to build a first (and maybe rough) geometric model, but with a few unknown parameters.
The data obtained from this optimization process were used to build a second-order stepwise linear regression model [ 37] (Supplementary Methods) to determine the relative importance of the individual drugs.
