-
Notifications
You must be signed in to change notification settings - Fork 0
samuelweisenthal/relative_sparsity
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
This is a readme for research code for the manuscript Relative Sparsity for Medical Decision Making Note that run.mc.R can be run locally or on a server, and that defaults here at set to server Note that there is a pdf that contains some preliminary data analysis for the real data, which was generated by an rmarkdown file Note that all file paths have been anonymized - so eg author names have been replaced with "anonymousAuthor" Files (alphabetically): analyze.sel.R analyzes Monte-Carlo results center.scale.R scales data. Option to center, not used is.R importance sampling (inverse probability weighting) functions mc.utils.R various helper functions for running Monte-Carlo experiments mcop This will just contain some intermediary results, in case it is useful to look back at them. Not really useful The main experiment results will be in a directory that is generated automatically by run.mc.R mimic.R Runs real data analysis modeling. Copied when run to save configurations mimic.prelim.analysis.Rmd Runs preliminary analysis for real data mimic.prelim.analysis.pdf Preliminary analysis run.mc.R Runs Monte-Carlo experiment. Copied when run to save configurations, and put in result directory Example_MC_ExperimentRes.Dir.n=22M=5plotsM=2K=2maxLam=2.00e+01minLam=1.00e-03usediff=0.001tau2=2 Just an example of a result directory generated by run.mc.R contents: logs empty because writes to stdout in local machine setting, but if on server (bh=1) logs will have the stdout masks.matrix object with all selections over MC datasets ors.av average coefficients, vn, etc over MC datasets. Is a list, which can be read and reprocessed eg for changing plots paramsel list with parameters for each MC dataset. ie seed, b0, etc. plots the main figure in the mansucript, lambda vs coefficients res.sel r objects for each individual MC dataset result unique.masks list of unique selections run.one.select.R Runs one selection sim.R Simulates (generates) data tag=mimicusediff=0.5,start.t3endStage3gammaselix=2minlam=1e-05maxlam=2000ncov=9use.dff=0.5.mimic.txt tag=mimicusediff=0.5,start.t3endStage3gammaselix=2minlam=1e-05maxlam=2000ncov=9use.dff=0.5resmimic.outer.resop These are just output files generated by mimic.R. Similar output files for Monte-Carlo are generated by run.mc.R and put in that results directory utils.R more general helper functions data_clean.py From Futoma 2020 paper. Creates episodes by processing csv files made by bash scripts in https://github.com/dtak/POPCORN-POMDP np_load.py Processes episodes created by data_clean.py so that they are in a form easily read by R. They will then be processed by mimic.R
About
No description, website, or topics provided.
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published