rdlearn - Safe Policy Learning under Regression Discontinuity Design with Multiple Cutoffs

Implements safe policy learning under regression discontinuity designs with multiple cutoffs, based on Zhang et al. (2022) <doi:10.48550/arXiv.2208.13323>. The learned cutoffs are guaranteed to perform no worse than the existing cutoffs in terms of overall outcomes. The 'rdlearn' package also includes features for visualizing the learned cutoffs relative to the baseline and conducting sensitivity analyses.