Split a dataset of out-of-sample predictions by estimate_type
Source: R/generate_yi_subsets.R
split_yi_subsets.RdReorganises the data from nesting based on individual analysis submissions, to nesting based on the type of estimate.
Value
A tibble of out-of-sample predictions subset by estimate_type and stored in the list-column data.
Details
This function is used to split a dataset of out-of-sample predictions by estimate_type, where out-of-sample predictions are stored in a list-column called back_transformed_data, with one data frame of data per analysis submission.
The estimate_type is derived from either the scenario and SurveyID columns in each back_transformed_data dataset.
The estimate_type is then used to nest the data by estimate_type in the list-column data.
Removes unnecessary data in .data: augmented_data and checks.
See also
This function is called in the wrapper function generate_yi_subsets() to split out-of-sample predictions by estimate_type across multiple ManyAnalyst datasets.