Description Usage Arguments Value
This function runs a simulation to compare LASSO, exposure based hdps, and bias based hdps as propensity score methods
1 2 3 4 5 |
simulationProfile |
simulationProfile object created by createCMDSimulationProfile function |
simulationRuns |
Number of simulations to run (1 simulation = reroll outcomes) |
trueEffectSize |
True effect size for exposure to simulate. If set to NULL keeps observed effect size from simulation profile |
outcomePrevalence |
Outcome prevalence to simulate; adjusts outcome baseline survival function to achieve. If null keeps observed |
hdpsFeatures |
TRUE = using HDPS features; FALSE = using FeatureExtraction features |
ignoreCensoring |
Ignore censoring altogether; sets censoring process baseline survival function to 1 |
studyPop |
Study population created by createStudyPopulation function |
confoundingScheme |
Type of unmeasured confounding to use for PS (0 = none; 1 = demographics; 2 = random proportion; 3 = demographics and random proportion) |
confoundingProportion |
Proportion of covariates to hide from propensity score as unmeasured confounding |
ignoreCensoringCovariates |
Ignore covariates effects on censoring process; only uses baseline function |
threads |
Number of parallel threads to use |
Returns the following:
the non-exposure oefficients used in the outcome generating model
coefficient for exposure effect
logRr, bound, and sd for LASSO estimate for each simulation
logRr, bound, and sd for exposure based hdps
logRr, bound, and sd for bias based hdps
auc for LASSO propensity score
auc for exposure based hdps propensity score
auc for bias based hdps propensity score
propensity scores for subjects in study population for LASSO
propensity scores for subjects in study population for exposure based hdps
propensity scores for subjects in study population for bias based hdps; propensity and preference scores are averaged over simulation runs
outcome prevalence of simulation
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.