Building predictive models with HED data using R


This talk will discuss the methods we use for building standardised readmission ratios in HED using Generalized Additive Models, the difficulties of handling HES data at scale and how we’ve transitioned from a single analyst with an R script to a team with robust modelling pipelines and metadata capture. We’ll touch on efficient methods for larger datasets, including using database storage and data.table, as well model checking and improvements.

Nov 11, 2020 —