10-24-2013 04:44 PM
I have fit a left truncated generalized pareto distribution using proc severity. I would like to use my parameter estimates to calculate expected values and to generate a random sample. Given that I do not observe any values below my truncation, and therefore cannot estimate the parameters of the unconditional CDF, I don't think I can use the CDF function with my estimated parameters to calculate the remaining area after subtracting the truncated portion. What should I use?