Batch Processing of Policy Trees and Related Visualisations
Source:R/margot_policy.R
margot_policy.Rd
Process a list of multi-arm causal forest results: generate policy-tree and decision-tree plots, Qini curves, and difference-gain summaries, optionally saving each to disk.
Arguments
- result_outcomes
List returned by
margot_multi_arm_causal_forest()
.- policy_tree_args
List of args for
margot_plot_policy_tree()
. Default:list()
.- decision_tree_args
List of args for
margot_plot_decision_tree()
. Default:list()
.- dpi
Resolution (dpi) for saved plots. Default: 600.
- width
Width (in inches) for saved plots. Default: 12.
- height
Height (in inches) for saved plots. Default: 12.
- save_plots
Logical; save plots to disk? Default: TRUE.
- output_dir
Directory for saving plots. Default:
here::here(push_mods)
.- spend
Numeric vector of spend levels for difference-gain summaries. Default:
c(0.2, 0.5)
.- label_mapping
Named list mapping variable names to display labels. Default: NULL.
- original_df
Optional data.frame of untransformed variables for axis annotations. Default: NULL.
- model_names
Character vector of model names to process; NULL = all. Default: NULL.
Value
A named list; each element corresponds to a model and contains:
- policy_tree_plot
ggplot of the policy-tree plot
- policy_tree_interpretation
character interpretation
- qini_plot
ggplot of the Qini curve
- decision_tree_visualisation
ggplot of the decision tree
- policy_combo_plot
ggplot of combined policy vs decision
- diff_gain_summaries
list of difference-gain summaries per spend