Stratification and Matching for Large Observational Data Sets

A pilot matching design to automatically stratify and match large datasets. The manual_stratify() function allows users to manually stratify a dataset based on categorical variables of interest, while the auto_stratify() function does automatically by allocating a held-aside (pilot) data set, fitting a prognostic score (see Hansen (2008) ) on the pilot set, and stratifying the data set based on prognostic score quantiles. The strata_match() function then does optimal matching of the data set in parallel within strata.