Does anybody have made an attempt already to code the method proposed in
Forecasting Using Maximum Entropy, Charles H. Whiteman, John C. Robertson, Ellis W. Tallman, Journal of Money, Credit, and Banking, vol 37, 2005, 383-401?
Or similar approaches? The method allows for imposing moment conditions (derived from external information) on the distribution of forecasts while changing the model generated forecast density as little as possible. The basic idea is to change the probabilities weights of the observations of the sampled density in such a way that the imposed moment conditions are fulfilled and the KLIC is minimized. The new weights (to be used in an adjusted density sampling) are obtained by solving a Lagragian optimisation problem.