question about gaussian_kern() parameters?

[ajdamico]

hi, hope all is well with you! we are reviewing some functions in our library(convey) to make sure our mathematics is sound and we're trying to reconcile why some of our at risk of poverty rate calculations are yielding very tiny differences from library(vardpoor) .. we have looked over the calculations in Osier's 2009 paper and we're curious if you might take a look at how linarpr implements the equation under section 2.2 on page 182 of the PDF below? the equation below By using the derivation rule (33), we obtain includes the threshold x=ARPT (M) but from our reading of linarpr, we're not sure if you're using x=MED (M) in this position instead? we've presented this as a pull request just to make it easier to highlight our question about the implementation but we definitely might be making a mistake or mis-reading the math on our end. we appreciate your time with this! -anthony and @guilhermejacob

https://www.researchgate.net/profile/Guillaume-Osier/publication/313562013_Variance_Estimation_for_Complex_Indicators_of_Poverty_and_Inequality_Using_Linearization_Techniques/links/597051ec4585158a48ffa05a/Variance-Estimation-for-Complex-Indicators-of-Poverty-and-Inequality-Using-Linearization-Techniques.pdf#page=17

[guilhermejacob]

For isntance, it appears the threshold is passed to the density function estimation in linpoormed, unlike linarpr.

vardpoor/vardpoor/R/linpoormed.R

Line 253 in 28ed111

u2 <- (thres_val - inco) / h