Generalized maximum entropy estimation of dynamic programming models with sample selection bias

From MaRDI portal





The author considers a class of censored Markov decision processes with decision variable censored at zero. The Lagrange multipliers technique is used to incorporate the restrictions. Maximum entropy principle is used in the selection of transition probabilities.











This page was built for publication: Generalized maximum entropy estimation of dynamic programming models with sample selection bias

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1424658)