Generates diagnostic plots for a list of robust state space models fit across a sequence of \(\lambda\) values. Two model attributes are plotted against \(\lambda\), each in its own panel, with a vertical dashed line indicating the model with the lowest BIC among those with fewer than 50% outliers.
Usage
# S3 method for class 'roams_SSM_list'
autoplot(object, attribute1 = "BIC", attribute2 = "prop_outlying", ...)Arguments
- object
An object of class
roams_SSM_listas returned byroams_SSMwhen multiple \(\lambda\) values are used.- attribute1
A character string indicating the first model attribute to plot on the top panel. Options include
"lambda","prop_outlying","BIC","loglik","RSS","iterations","value", and"counts". Defaults to"BIC".- attribute2
A character string indicating the second model attribute to plot on the bottom panel. Uses the same options as
attribute1. Defaults to"prop_outlying".- ...
Other arguments passed to specific methods. Not used in this method.
Value
A ggplot object arranged using the patchwork package, showing the specified attributes plotted against \(\lambda\) in two vertically stacked panels.
Details
In each panel, a red dashed vertical line indicates the model with the lowest BIC among those with fewer than 50% outlying time points, serving as a heuristic for robust model selection.