plot.bdt: Plotting of the results of a call to the bdt function

In Yan2020729/bdt1: Bootstrap Diagnostic Tool

Description

This functions is methods for class bdt objects

Usage

  ## S3 method for class 'bdt'
plot(x, xlab = NULL, ylab = "Bias", outlierSize= 0.4, notch = FALSE, pointsize = 0.7, pointcol = "blue", labelsize = 0.35, labelcol = "red", linetype = "dotted", linecol = "darkred", ...)

Arguments

x	an object of class bdt.
xlab	x-asix label. Default value is NULL.
ylab	x-asix label. Default is "Bias".
outlierSize	the point size of outlier. Default is 0.4
notch	notch for boxplot. Default value is FALSE.
pointsize	the size of point. Default is 0.7.
pointcol	the color of points. Default is blue.
labelsize	the size of the label of coverage rates. Default is 0.35.
labelcol	the color of the label of coverage rates. Default is red.
linetype	the type of reference line. Default is dotted line.
linecol	the color of reference line. Default is darkred.
...	graphics parameters to be passed to the plotting routines.

Value

boxplot_bias_ATE	the boxplots and points of bias of the three estimated ate and coverage rates where y-axis represents the bias of estimates, x-axis represents estimates: AIPTW, TMLE and coverage rates are marked as red in corresponding boxes
densityplot_ps	the 2 x 2 density plots (A, B, C, D) of the log of the estimated weights where A refers to treatment A=1 in subgroups with A=1, B refers to treatment A=0 in subgroups with A=0, C refers to treatment A=1 for all subjects and D refers to treatment A=0 for all subjects

References

1. Bahamyirou A, Blais L, Forget A, Schnitzer ME. (2019), Understanding and diagnosing the potential for bias when using machine learning methods with doubly robust causal estimators. Statistical methods in medical research, 28(6), 1637-50.

See Also

bdt summary.bdt

Yan2020729/bdt1 documentation built on March 24, 2021, 8:58 p.m.