kr_cate: Estimates non-paramteric CATEs using kernel regression as...

In MCKnaus/causalDML: Causal Double Machine Learning

Estimates non-paramteric CATEs using kernel regression as proposed by Fan et al. (2019) and Zimmert & Lechner (2019).

Description

Estimates non-paramteric CATEs using kernel regression as proposed by Fan et al. (2019) and Zimmert & Lechner (2019).

Usage

kr_cate(delta, z, bw_factor = 0.9)

Arguments

delta	Vector of doubly robust ATE score. E.g obtained as one column of ATE_dml$delta from ATE_dml.
z	Heterogeneity variable(s) vector, matrix or data.frame.
bw_factor	Factor by which cross-validated is multiplied. Default is undersmoothing with factor 0.9 as recommended by Zimmert & Lechner (2019).

Value

kr_cate object:

model	npqregression object (seenpreg) of the kernel regression
fit	Fitted values of the kernel regression
bw	Cross-validated bandwidth (not scaled)
ate	Average treatment effect

References

• Fan, Q., Hsu, Y.-C., Lieli, R. P., & Zhang, Y. (2019). Estimation of conditional average treatment effects with high-dimensional data. arXiv preprint arXiv:1908.02399. http://arxiv.org/abs/1908.02399

• Zimmert, M., & Lechner, M. (2019). Nonparametric estimation of causal heterogeneity under high-dimensional confounding. arXiv preprint arXiv:1908.02399. http://arxiv.org/abs/1908.08779

MCKnaus/causalDML documentation built on June 11, 2025, 12:30 a.m.