plotci: Plot Wald Confidence Intervals

In andrew-mcinerney/statnn: Interpreting Neural Networks From a Statistical Perspective

Plot Wald Confidence Intervals

Description

plotci It is designed for an inspection of the confidence intervals for the weights for objects of class interpretnn.

Usage

plotci(
  object,
  alpha = 0.05,
  which = c(1L:object$n_inputs),
  which_params = c(1L:object$n_nodes),
  colour = 1,
  ask = prod(graphics::par("mfcol")) < length(which) && grDevices::dev.interactive(),
  caption = lapply(1:ncol(object$X), function(iter) {
    
    paste0("Confidence Intervals for ", colnames(object$X)[iter])
 }),
  ...
)

Arguments

object	an object of class interpretnnet.
alpha	significane level.
which	index of plots to be displayed.
which_params	index of weights to be displayed.
colour	colour of confidence intervals.
ask	ask before displaying each plot.
caption	caption for each plot.
...	arguments to be passed to methods, such as graphical parameters (see par).

Value

plot of weights and their significance

andrew-mcinerney/statnn documentation built on June 30, 2024, 4:09 p.m.