Software‎ > ‎

Model-free covariate selection

CovSel

Model-free selection of covariates under unconfoundedness for situations where the parameter of interest is an average causal effect. This package is based on model-free backward elimination algorithms proposed in de Luna, Waernbaum and Richardson (2011, Biometrika). Marginal co-ordinate hypothesis testing is used in situations where all covariates are continuous while kernel-based smoothing appropriate for mixed data is used otherwise.

Link to CRAN-R.


Model-free selection of covariates in high dimensions under unconfoundedness for situations where the parameter of interest is an average causal effect. This package is based on model-free backward elimination algorithms proposed in de Luna, Waernbaum and Richardson (2011) <doi:10.1093/biomet/asr041> and VanderWeele and Shpitser (2011) <doi:10.1111/j.1541-0420.2011.01619.x>. Confounder selection can be performed via either Markov/Bayesian networks, random forests or LASSO

Link to CRAN-R


References:

SLIDES from invited talk at Joint Statistical Meeting in Seattle, 2015.

Häggström, J., Persson, E., Waernbaum, I., & de Luna, X. (2015). CovSel: An R Package for Covariate Selection When Estimating Average Causal Effects. Journal of Statistical Software, 68(1), 1 - 20. doi:http://dx.doi.org/10.18637/jss.v068.i01

Persson, E., Häggström, J., Waernbaum, I., & de Luna, X. (2016). Data-driven algorithms for dimension reduction in causal inferenceComputational Statistics & Data Analysis, 105, 280-292. Preprint. arXiv:1309.4054

de Luna, X., Waernbaum, I., and Richardson, T.  (2011)  Covariate selection for the non-parametric estimation of an average treatment effect, Biometrika, 98, 861-875.

Häggström, J. (in press). Data-driven Confounder Selection via Markov and Bayesian Networks (with discussion). Biometrics, to appear. Preprint: arXiv:1604.07212
Comments