SL.wang: Quantile regression method of Wang and Zhou (2009)

In wuziyueemory/twostageSL: Two-Stage Super Learner Prediction Function

Description

This function implements the estimator of Wang and Zhou (2010). The estimator estimates the conditional mean costs by modeling the conditional quantiles of a transformed cost. Using the equivariance property of quantiles to monotone transformations, the quantile estimators may be transformed back to the conditional mean cost on the original scale.

Usage

SL.wang(
  Y,
  X,
  newX,
  family = gaussian(),
  obsWeights = rep(1, length(Y)),
  g = "log",
  m = length(Y),
  c = 0.2,
  b = 0.05,
  ...
)

Arguments

Y	A numeric outcome variable
X	A data.frame of covariates constituting the training sample
newX	A data.frame with the same column names and format as X constituting the validation sample.
family	Gaussian only
obsWeights	Observation-level weights (not currently used)
g	Transformation to apply to Y before quantile regression is used. Choices are "log" or "sqrt"
m	Number of quantiles to compute
c	Constant used to determine truncation level for transforming quantiles to conditional mean
b	Constant used to determine truncation level for transforming quantiles to conditional mean
...	Other arguments (not currently used)

Value

pred

Predicted outcomes based on predictors in newX

fit

A list with named entries object (the fitted rq regression object), alpha (the controlled level of trimming based), and g_inv (the inverse function of the inputted g)

References

Wang HJ, Zhou X (2010). “Estimation of the retransformed conditional mean in health care cost studies.” Biometrika, 97(1), 147–158.

Examples

# load cost data
data(cost_data)
# fit wang model
fit_wang <- SL.wang(Y = cost_data$totalcost, X = cost_data[, c("female", "race")],
                    newX = cost_data[, c("female", "race")])

wuziyueemory/twostageSL documentation built on Oct. 19, 2020, 3:45 p.m.