Columbia University, Mailman School of Public Health. Environmental Mixtures Workshop 2018/2019. Introduction to multiple techniques to analyze exposure to mixtures in environmental health.
Within this repository you will find the code materials used during the Mixtures Workshop 2018/2019. To help you navigate the repository, below is a description of the repository organization.
The repository is organized into two main folders "Unsupervised" and "Supervised" which contained the materials for the unsupervised and supervised methods correspondingly. The Unsupervised folder is subdivide into three folders: PCA, FA, and Clustering. And the Supervised folder is subdivide into: WQS, Variable Selection, and BKMR. All the materials for each respective method are within its folder. Note, the rmd file titled with the method's name contains the main code.
In addition to the Unsupervised and Supervised folders you will find the following folders:
- Data: contains the data and the data dictionary.
- Mitro_et_al_materials: contains the code and data used by Mitro et al.
- workshop_paper_figures: contains the figures include in the Gibson et al. paper.
WARNING: This repo was created under R version 3.5.3. Results will differ with newer R versions, as R version 3.6.0 changed the default method for generating from a discrete uniform distribution (used in sample()). Additionally, version 3.1-3 of the grpreg package should be used to recreate published results (here).