Fisher's identity is useful to use in maximum-likelihood parameter estimation problems. In this post, I give its proof. The main reference is Douc, Moulines, Stoffer; Nonlinear time series theory, methods and applications.
Subscribe to:
Posts (Atom)
