This function performers an approximated Bayesian cross-validation for a bcmeta object and specially designed diagnostics to detect the existence of a biased component.
# S3 method for bcmixmeta
diagnostic(
object,
post.p.value.cut = 0.05,
study.names = NULL,
size.forest = 0.4,
lwd.forest = 0.2,
shape.forest = 23,
bias.plot = TRUE,
cross.val.plot = FALSE,
level = c(0.5, 0.75, 0.95),
x.lim = c(0, 1),
y.lim = c(0, 10),
x.lab = "P(Bias)",
y.lab = "Mean Bias",
title.plot = paste("Bias Diagnostics Contours (50%, 75% and 95%)"),
kde2d.n = 25,
marginals = TRUE,
bin.hist = 30,
color.line = "black",
color.hist = "white",
color.data.points = "black",
alpha.data.points = 0.1,
S = 5000,
...
)The object generated by the function b3lmeta.
Posterior p-value cut point to assess outliers.
Character vector containing names of the studies used.
Size of the center symbol mark in the forest-plot lines
Thickness of the lines in the forest-plot
Type of symbol for the center mark in the forest-plot lines
Display the bias plot. The default is TRUE.
Display the cross validation plot. The default is FALSE.
Vector with the probability levels of the contour plot. The default values are: 0.5, 0.75, and 0.95.
Numeric vector of length 2 specifying the x-axis limits.
Numeric vector of length 2 specifying the y-axis limits.
Text with the label of the x-axis.
Text with the label of the y-axis.
Text for setting a title in the bias plot.
The number of grid points in each direction for the non-parametric density estimation. The default is 25.
If TRUE the marginal histograms of the posteriors are added to the plot.
The number of bins in for the histograms. The default value is 30.
The color of the contour lines. The default is "black.
The color of the histogram bars. The default is "white".
The color of the data points. The default is "black".
Transparency of the data points.
The number of sample values from the joint posterior distribution used to approximate the contours. The default is S=5000.
...