Link Search Menu Expand Document

Welcome to the AHB R Package Documentation!

View us on GitHub


AHB is an R package for performing matching for causal inference on observational datasets containing discrete, continuous, or mixed covariates. AHB matches treatment and control units in unit-specific, hyper-box-shaped regions of the covariate space. This package offers two implementations of the AHB algorithm, one which calculates regions as the solution to a mixed integer program (MIP) and another which uses a fast approximation algorithm.

Contact

Please reach out to let our team know if you’re using this, or if you have any questions! Email us at almost-matching-exactly@duke.edu.