Run the random walk Metropolis algorithm with a normal or a Student-t proposal density, the slice sampler, the fixed blocking RWM sampler with a normal or a Student-t proposal density, the random blocking RWM sampler with a normal or a Student-t proposal density, the Sequential Monte Carlo with likelihood or data tempering, or importance sampling based on the MitISEM algorithm to sample from the posterior distribution of the parameters of the DSGE model. The number of posterior draws, the number of burn-in draws, the number of parallel sampling chains, etc., are determined in the posterior sampling frame on the Options tab.
The parameterization of the proposal distribution for the random walk Metropolis is also determined from your settings in the posterior sampling frame. Note that YADA always samples directly from the transformed parameters. The program then converts each draw to the original form.
Once the sampler has finished it will compute statistics for the estimated parameters, such as the posterior mean and median. Optionally, it will also compute the marginal likelihood, depending on your selections in the DSGE posterior sampling frame on the Settings tab. If you have selected a marginal likelihood algorithm it will run this for the whole post burn-in sample. In addition, if you have selected to compute the marginal likelihood sequentially, this will also be performed. However, such sequential estimation of the marginal likelihood can also be performed from the View menu.
Finally, YADA writes a summary of the results to file and displays this file. If you wish to review your results later on you can also access them via the posterior sampling summary function on the View menu.
If you wish to view graphs of the marginal posterior densities you can do so via the posterior densities function on the View menu.
This function is also available on the toolbar.
Icon: |
Additional Information
• | A details discussion of parameter transformations is found in Section 6 of the YADA Manual. |
• | A detailed description of the random walk Metropolis related algorithms can be found in Section 8.1 of the YADA Manual. |
• | A detailed description of the slice sampler can be located in Section 8.2 of the YADA Manual. |
• | A detailed description of the Sequential Monte Carlo with likelihood or data tempering samplers are found in Section 8.4 of the YADA Manual. |
• | A detailed description of the Importance Sampler based on the MitISEM algorithm for estimating the importance (candidate) density is provided in Section 8.5 of the YADA Manual. |
Page url: http://www.texlips.net/yada/help/index.html?posterior_sampling2.htm