This code is trying to replicate the work of Cabras and Castellanos (2011) (A Bayesian approach for estimating extreme quantiles under a semiparametric mixture model) using MATLAB language.
`DanishData.txt': Danish fire loss data
'data.xlsx': the other data
TestCase.m: main file
LogLikLiHood.m: for computing the log likelihood, including the following sub functions
LindseyMethod.m: for implementing the Lindsey method in the paperOrthogonalPoly.m: for computing the orthogonal polynomials, corresponding to the “poly()” in ROrthogonalPredict.m: for predicting the value of orthogonal polynomial for new data, corresponding to the “predict()” in R
posterior.m: for computing the posterior probability. There are some adjustments, which may cause confusion.
gwmcmc.m: for implementing the Markov Chain Monte Carlo simulation. The original code is written by Aslak Grinsted and can be found here.