Imputation of LCFS onto FRS

[nikhilwoodruff]

After the LCFS openfisca-uk-data dataset is complete, we should plan out how best to impute LCFS model variables onto the FRS dataset. It looks like the best way to do this is using the random forest-based imputation with Beta distribution matching of aggregates from rf_impute in synthimpute.

[MaxGhenis]

For this project, we'd only need to impute carbon emissions right?

[nikhilwoodruff]

Yeah, I think so. Though an alternative might be imputing the expenditure on the different categories, and calculating carbon emissions within openfisca using parameters - an advantage of this is that it's then a lot simpler to later come back to when we add VAT, but probably overcomplicated for this project?