The goal of this work is to build a predictive model for lactate-phenotype discordance. This is ongoing work and some of the files are subject to change. The portion of this project written up to 5/2018 represents my practcium work for the completion of a Masters of Public Health in Quantitiave Methods at Harvard Chan School.
This repository is organized as follows:
- Extraction: contains a notebook and requisite scripts for extacting cohort data from the eICU database
- EDA: contains a notebook with an exploratory analysis of the cohort as well as a notebook examining the association between lactate discordance and mortality. (Note on mortality: As this was examined after the initial extract, this file performs a database pull to attain mortality status.)
- Classifier: contains a notebook which documents the construction of a classifier