Weighted Median bootstrap

Question

Currently I am obtaining 95% CI on the median by using this

x<-rnorm(100)
        bootmed = apply(matrix(sample(x, rep=TRUE, 10^4*length(x)), nrow=10^4), 1, median)
    quantile(bootmed, c(.025, 0.975))[1]->a1
    quantile(bootmed, c(.025, 0.975))[2]->a2

the problem now is that i need to do this for a weighted Median. I use the weightedMedian function in the matrixStats package: matrixStats::weightedMedian

so I do not only have x with the numbers but also y with the weights (runif(100)) - so now I calculate

weightedMedian(x,runif(100))

but how does the equivalent for the above bootstrap go?

Answer 1

The ... argument of the apply() function allows you to pass optional arguments to your function. Alternatively, you can also define a custom function. Assuming that your median weights are identical for all rows, the answer to your question looks like this:

x <- rnorm(100)
y <- rnorm(100)
M <- matrix(sample(x, rep=TRUE, 10^4*length(x)), nrow=10^4)

bootmed = apply(M, 1, weightedMedian, w=y)
# Or, alternatively..
bootmed = apply(M, 1, function(x) weightedMedian(x,y))