Matlab fminunc calculate standard errors MLE -


i have problem when trying calculate standard errors of estimates fminunc. estimation technique maximum likelihood estimation. i've tried 2 ways below, both failed:

  1. the hessian matrix fminunc results non-invertible, can't standard error hessian matrix.

  2. so turned standard errors using opg(outer-product-of-gradient) method. however, gradient provided failed pass derivative-check.

does know whether there other ways can standard error? appreciated!

this may useful: http://gking.harvard.edu/files/gking/files/help.pdf

i've skimmed , sense can't variance matrix non-invertible hessian matrix. doesn't exist. reference above says standard response is:
a) more data or
b) use less complicated model.
offers rescue methods if neither of option. haven't gotten head around methods, sorry incomplete.


Comments

Popular posts from this blog

javascript - JS causing window size to be bigger than necessary - Dropdown bug -

How to mention the localhost in android -

php - Calling a template part from a post -