-2

i'm new here and i am super desperate so i really hope anyone of you can help me.... i have a sample of random data x_1....x_n and i want to fit a truncated pareto distribution to the data.... to fit a generalized pareto distribution is super easy and i have already done that. I calculated the shape and scale parameters with a matlab routine. But for the truncated pareto distribution i can't seem to find a routine to calculate the parameters i need... Does anybody have an idea how to do it?

Thanks in advance!

keep_isi
  • 1
  • 1
  • Try https://www.mathworks.com/help/stats/examples/fitting-custom-univariate-distributions.html – Dandan Jan 12 '18 at 00:14

1 Answers1

0

You can use Markov-Chain-Monte-Carlo simulations to do Bayesian Inference to get the most likely parameters of your generalized pareto distribution for the given data. Or you stay with the maximum likelihood method. Your problem can be solved in many ways. But if you want to apply MLE you actually just need to search for a maximum. You could do it with fminsearch() http://de.mathworks.com/help/optim/ug/fminsearch.html

For this you just need to define another function in a separate m-file which computes your Likelihood or Log-Likehood for a given set of parameters of your truncated pareto distribution. fminsearch now returns you the optimal parameters according to this likelihood. Is this the kind of routine you are looking for?

v.tralala
  • 1,444
  • 3
  • 18
  • 39