Released on 21/01/2022:
• | Added predictive likelihood estimation routines to the adaptive learning tools for marginal and joint predictions using the posterior draws. It also supports back- and nowcasts for the original data and marginalized predictive likelihoods if the historical data has a so called ragged edge. |
• | Added unconditional predictive distributions, PIT and CRPS/ES calculations for the posterior draws to the adaptive learning tools. |
• | Added the import prior data option to the learning menu options. |
• | Added import of initial parameter values and the inverse Hessian from the posterior mode estimates of another model. This is mainly intended when estimating models recursively and where, for example, past estimates can serve as potentially good initial values. |
• | Added sequential posterior mode estimation, sequential posterior sampling, sequential estimation of the predictive likelihood, sequential estimation of marginal predictive moments, sequential marginal likelihood estimation and sequential estimation of the probability integral transforms (PITs) for DSGE models with adaptive learning. Model sequences can be selected and the tools run via items on the Model sequence menu-item on the File menu. |
Page url:
http://www.texlips.net/yada/help/index.html?4_90.htm