pif.confidence.bootstrap: Pivotal Boostrap Confidence Intervals for Potential Impact...

In INSP-RH/pif: Potential Impact Fraction and Population Attributable Fraction for Cross-Sectional Data

Description

Estimates a 1 - alpha pivotal confidence interval for the potential impact fraction pif using a boostrap approximation.

Usage

pif.confidence.bootstrap(X, thetahat, thetavar, rr, cft = NA,
  weights = rep(1/nrow(as.matrix(X)), nrow(as.matrix(X))),
  method = c("empirical", "kernel"), nboost = 10000, adjust = 1,
  n = 512, confidence = 95, ktype = c("gaussian", "epanechnikov",
  "rectangular", "triangular", "biweight", "cosine", "optcosine"),
  bw = c("SJ", "nrd0", "nrd", "ucv", "bcv"), check_exposure = TRUE,
  check_rr = TRUE, check_integrals = TRUE, check_thetas = TRUE,
  is_paf = FALSE)

Arguments

X	Random sample (vector or matrix) which includes exposure and covariates.
thetahat	Maximum Likelihood estimator (vector or matrix) of theta for the Relative Risk function.
thetavar	Estimator of variance of thetahat
rr	Function for Relative Risk which uses parameter theta. The order of the parameters shound be rr(X, theta). **Optional**
cft	Function cft(X) for counterfactual. Leave empty for the Population Attributable Fraction paf where counterfactual is 0 exposure.
weights	Normalized survey weights for the sample X.
method	Either empirical (default), or kernel
nboost	Number of samples in Bootstrap
adjust	Adjust bandwith parameter from density (for kernel method) from density.
n	Number of equally spaced points at which the density (for kernel method) is to be estimated (see density).
confidence	Concidence level (0 to 100) default = 95 %
ktype	kernel type: "gaussian", "epanechnikov", "rectangular", "triangular", "biweight", "cosine", "optcosine" (for kernel method). Additional information on kernels in density
bw	Smoothing bandwith parameter from density (for kernel method) from density. Default "SJ".
check_exposure	Check that exposure X is positive and numeric
check_rr	Check that Relative Risk function rr equals 1 when evaluated at 0
check_integrals	Check that counterfactual and relative risk's expected values are well defined for this scenario
check_thetas	Checks that theta parameters are correctly inputed
is_paf	Boolean forcing evaluation of paf

Value

pif Estimate of Potential Impact Fraction

Author(s)

Rodrigo Zepeda-Tello [e-mail]

Dalia Camacho-García-Formentí [e-mail]

Examples

#Example 1: Exponential Relative Risk
#----------------------------------------
set.seed(18427)
X        <- rnorm(100,5,1)
thetahat <- 0.4
thetavar <- 0.1
pif.confidence.bootstrap(X, thetahat, thetavar, function(X, theta){exp(theta*X)}, 
                         nboost = 100) #nboost small only for example purposes

#This also works with kernel method
pif.confidence.bootstrap(X, thetahat, thetavar, function(X, theta){exp(theta*X)}, 
                         nboost = 100, method = "kernel") 

#Example 2: Multivariate example
#----------------------------------------
## Not run: 
set.seed(18427)
X1 <- rnorm(100, 1, 0.05)
X2 <- rnorm(100, 1, 0.05)
X  <- as.matrix(cbind(X1,X2))
thetahat <- c(2, 0.03)
thetavar <- matrix(c(0.1, 0, 0, 0.05), byrow = TRUE, nrow = 2)
rr        <- function(X, theta){
  .X <- as.matrix(X, ncol = 2)
  exp(theta[1]*.X[,1] + theta[2]*.X[,2])
}
cft <- function(X){0.5*X}#' cft <- function(X){0.95*X}
pif.confidence.bootstrap(X, thetahat, thetavar, rr, cft) 

## End(Not run)

INSP-RH/pif documentation built on May 7, 2019, 6:01 a.m.