get_piahat: Estimation of regression function E(A|X)

In matteobonvini/sensitivitypuc: Sensitivity Analysis via the Proportion of Unmeasured Confounding (PUC)

Description

get_piahat estimates the regression function E(A|X) using the SuperLearner.

Usage

get_piahat(a, x, newx, family = binomial(), trunc_tol = 0.05, sl.lib)

Arguments

a	nx1 vector of treatments
x	nxp data.frame of covariates
newx	qxp data.frame of values that the regression estimates are evaluated at
family	family specifying the error distribution for treatment regression, currently only binomial() supported. Default is binomial(). Link should not be specified.
trunc_tol	amount of tolerance allowed for truncating the propensity score in [tol, 1 - tol]. Default is 0.05.
sl.lib	character vector specifying which libraries to use for the SL.

Value

A list containing estimates of E(A|X):

testvals	nrow(newx)-dimensional vector containing estimates of regression functions (computed using points x) evaluated at test points newx.
trainvals	nrow(x)-dimensional vector containing estimates of regression functions (computed using points x) evaluated at train points x.

Details

If the SuperLearner returns an error, a GLM is fitted instead. In this case the user is suggested to choose some other method of estimation and then pass the estimates as arguments to other functions. Sometimes SuperLearner returns warnings messages.

References

Van der Laan, M. J., Polley, E. C., & Hubbard, A. E. (2007). Super learner. Statistical applications in genetics and molecular biology, 6(1).

See Also

do_crossfit and get_muahat.

Examples

n <- 500
x <- data.frame(x1 = rnorm(n), x2 = runif(n))
newx <- data.frame(x1 = rnorm(20), x2 = runif(20))
a <- rbinom(n, 1, pnorm(x$x1))
fits <- get_piahat(a, x, newx, family = binomial(), 
                   sl.lib = c("SL.mean", "SL.glm", "SL.gam"))
head(fits$testvals)

matteobonvini/sensitivitypuc documentation built on Dec. 9, 2020, 2:24 a.m.