Dirichlet regression of alpha values

I want to regress alpha values for a prospective dirichlet distribution. In my case, I have an excel file vector of 42 values of x for example (arranged in ascending order of magnitude). I want to regress 42 alphas so that the resultant 42 x's can fit my excel file x's well, when in ascending order. Note that I eventually will generate 200 instances of the dirichlet distribution, arrange each instance in ascending order and take avaerage of each xi, and that should fit my initial vector of x provided in excel. I am a novice in R and statistics, so please excuse my language.

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.