treynkens / reins Goto Github PK

Hey @TReynkens,

I've noticed in practice that some data sets and splicing point settings can lead to an invalid estimate of the upper truncation point for the spliced mixed Erlang and truncated Pareto distribution.

Here demonstrated with some generated data:

library(ReIns)

# Pareto random sample
set.seed(123)
X <- rpareto(500, shape = 2)

# Splice ME and Pareto
splicefit <- SpliceFitPareto(X, const = 0.9, EVTtruncation = TRUE, M = 5, s = 1)
#> Error in if (endpoint <= scale) {: missing value where TRUE/FALSE needed

^{Created on 2023-11-02 with reprex v2.0.2}

Rcpp will by default define STRICT_R_HEADERS in September 2019, see email Dirk Eddelbuettel on 18/09/2018 and RcppCore/Rcpp#898.
For ReIns, DBL_MAX is causing a problem.

Hi, I am impatient to see this book published... Btw, I was wondering wether there are any plan to implement a function for the Gertensgarbe-Werner estimator. Once I was given a matlab code that I attach
GartensGarbe.m.txt that migth be helpful