dualbounds.dist_reg.QuantileDistReg.predict¶

QuantileDistReg.predict(X: array, W: array, Z: array | None = None)[source]¶

Predicts the conditional law of the outcome.

Parameters:¶

X : np.array¶

(n, p)-shaped array of covariates.

W : np.array¶

Optional n-length array of binary treatment indicators.

Z : np.array¶

Optional n-length array of binary instruments for the instrumental variables setting.

Returns:¶

y_dists – batched scipy distribution of shape (n,) where the ith distribution is the conditional law of \(Y_i | X_i, W_i, Z_i\) (without conditioning on \(Z_i\) if Z is not provided).

Return type:¶

stats.rv_continuous / stats.rv_discrete