plotBias: plotBias

In benkeser/haltmle.sim: Simulation study for HAL-TMLE

Usage

plotBias(result, estNames, strat, truthName = "truth",
  pointOptions = list(pch = 1:length(estNames)), estLabels = estNames,
  legendOptions = list(x = "topleft", legend = estLabels), ylim = NULL,
  pctTruth = TRUE, summary = list(fn = "quantile", options = list(probs =
  0.5)), ...)

Arguments

result	The result data frame
estNames	The names of the estimators to plot
strat	The name of the variable on which to stratify
truthName	The column name containing the true ATE
pointOptions	A list of arguments to be passed to points
estLabels	Vector of names to label estimators in legend (default is to use estNames)
legendOptions	A list of arguments passed to legend. pointOptions are added automatically
ylim	y-axis limits for the plot. If null, looks for range and adds 5 \item pctTruthReport error as absolute percent of true ATE? If FALSE, error is reported as absolute units of information (i.e., inverse of variance of EIF) \item summaryA list with two entries named "fn" and "options". This is the summary function used to summarize percent difference. Defaults to median (i.e., function = quantile, options = list(probs = 0.5)). Assumes that the function will play nicely with the unnamed argument passed in \item ...Additional arguments passed to plot

Function to plot the median percent bias for a given set of estimators stratified by a given set of covariates

benkeser/haltmle.sim documentation built on May 12, 2019, noon