Docs

std.causal.crossfit

Cross-fitting and double machine-learning (DML) estimators.

(import std.causal.crossfit)

Symbol	Description
`(crossfit:k-folds data k seed)`	Partition `data` into `k` folds.
`(crossfit:nuisance data folds spec target cols)`	Cross-fit a single nuisance regression.
`(crossfit:nuisance-arm data folds spec y x arm cols)`	Cross-fit per treatment arm.
`(crossfit:dml-plr data y x cols . opts)`	Partially-linear DML estimator.
`(crossfit:dml-irm data y x cols . opts)`	Interactive regression DML estimator.
`(crossfit:influence-se psi)`	Standard error from influence values.
`(crossfit:dml-ci estimate se [alpha])`	Confidence interval from DML output.

Options include 'folds, 'seed, 'outcome-spec, 'propensity-spec.