Serving the Quantitative Finance Community
distribution fit
HiAssume we have at disposal historical returns for HF (at daily freq) and that we aim at finding the best fit for the distribution.That is to find the distribution class (Student-t, Levy...) and the parameters (dof...).What are the relevant steps to optimize this problem?1.Interpolate/extraploate the pdf (what algorithm as spline won't do the job correctly)?2.LSE for finding the parameters?Thks for your hints
distribution fit
Not sure what you mean by "Interpolate/extraploate the pdf". But in general, you would want to do maximum likelihood estimation, using some sort of algorithm to traverse parameter space. Be sure to get good initial estimates though. This type of stuff is more art than science though.
distribution fit
There are good answers from fermion and 2 in the last days:"find a parametric model to fit data" and "accurate extraploation methods for few points"