build_covariate: Build covariance matrix of the covariates

In ygeunkim/propensityml: Propensity score weighting using machine learning methods

Description

builds covariance matrix of the onfounders, exposure predictors, and outcome predictors

Usage

build_covariate(
  sig = rep(1L, 10),
  confounder_exposure = c(0.2, 0, 0, 0, 0.9, 0, 0, 0, 0, 0, 0, 0),
  confounder_outcome = c(0, 0, 0, 0, 0, 0, 0.2, 0, 0, 0, 0.9, 0),
  exposure_outcome = rep(0L, 9),
  value_cor = TRUE
)

Arguments

sig	variance of each covariate
confounder_exposure	covariance between confounder and exposure predictors, vector in order of column
confounder_outcome	covariance between confounder and outcome predictors, vector in order of column
exposure_outcome	covariance between exposure predictors and outcome predictors, vector in order of column
value_cor	are given values (except sig) are correlation? (default = TRUE)

Details

This function builds covariance matrix of the covariates, which are confounders (w1, w2, w3, w4), exposure predictors (w5, w6, w7), and outcome predictors(w8, w9, w10). In each group, there is no correlation. On the other hand, you can specify picewise correlation between e.g. confounder-exposure predictor.

References

Setoguchi, S., Schneeweiss, S., Brookhart, M. A., Glynn, R. J., & Cook, E. F. (2008). Evaluating uses of data mining techniques in propensity score estimation: a simulation study. Pharmacoepidemiology and Drug Safety, 17(6), 546–555 https://doi.org/10.1002/pds.1555

See Also

sim_outcome

ygeunkim/propensityml documentation built on Jan. 1, 2021, 1:44 p.m.