Hello!
I am trying to develop a weighted quantile sum regression-based index and have been using the estimate.wqs function (https://rdrr.io/cran/miWQS/man/estimate.wqs.html). I am running into two issues, and I have not been able to find a solution. I have a dataset of 108 observations and am including 8 domains as the independent variables to be weighted. Here is my code (modeled directly after the above documentation), and my issues/questions are described below it:
if( requireNamespace("wqs", quietly = TRUE) ){
library(miWQS)
set.seed(23456)
Wa=estimate.wqs(y=df.2$LEB, X=df.2[,2:9],
B = 100, verbose=(TRUE))
Wa
}else{
message("you need to install the package wqs for this example")
}
Issue 1: Bootstrap convergence error
I received this error message: "5 bootstrap regression estimates in the training sets have not converged. Proceed results with caution." A colleague suggested I adjust the iterations within a given bootstrap, but I am not sure how to do that within this code. Does anyone know, or are there any other thoughts on how to ensure that all boostrap regressions converge?
Issue 2: Setting minimum weights
I would like to set a minimum weight of 0.05 for each of the 8 variables. How can I go about doing that?
Thank you for your help!