In mac-theobio/effects: isolated confidence intervals and bias-corrected predictor effects

library(ggplot2)

library(vareffects); varefftheme()

set.seed(9991)
load("variable_predictions_objs.rda")
load("binom_correlated_preds.rda")

We want:

[ E(y|{X_f, \bar{X}_{{n}}}) ]

Challenges:

\begin{enumerate} \item choice of representative values of \emph{focal} predictor(s) and appropriate ``model center" for \emph{non-focal} predictors especially in multivariate models \item propagation of uncertainty ([how] can we incorporate uncertainty in nonlinear components of (G)LMMs? Should we exclude variation in non-focal parameters?) \item bias in the expected mean prediction induced by the nonlinear transformation of the response variable (especially in GL(M)M) \end{enumerate}

print(pred_age_cont_plots)

print(pred_bin_corr_plots)

mac-theobio/effects documentation built on July 6, 2023, 4:19 a.m.