Closed timmens closed 4 years ago
In commit (61ed772f39e) I add a prediction method for multiple new observations (predict_causaltree
), which is based on the internal function _predict_row_causaltree
that predicts the estimated treatment effect for a single observation.
In commit (7aa3f9a91b92f) I add tests for both functions.
For a fitted (causal) tree --stored as a pandas DataFrame-- implement a prediction function which finds the leaf in which new observations fall and returns the treatment effect in the given leaf.