Differences between revisions 8 and 9

New Estimation Interface

Suggestions for a simplification of the estimation interface in Dynare and breaking very general commands with lots of options in several commands with more targeted scope.

set_time

dataset

estimated_params

Rely on options names rather than argument position in the command line. This new interface would replace estimated_params_bounds and estimated_params_init. The following option names would be necessary

For the declaration of the priors
- shape, name of the prior distribution.
- mean, mean of the prior distribution.
- stdev, standard deviation of the prior distribution.
- variance, variance of the prior distribution.
- mode, mode of the prior distribution.
- interval, specification of the prior distribution from the cumulative distribution function defining an interval covering an arbitrary percentage of the prior mass. See below for an example.
- shift, to shift the domain a distribution. See below for an example (useless if the domain of the distribution is the set of real numbers).
- domain, specify the domain of distributions with bounded support (uniform and beta).
For optimization options
- boundaries, enforce the optimizer to look for a mode in a specified interval.
- initial, initial value for the optimizer.
For the mcmc options
- jscale, scale paramemeter for the jumping distribution specific to a parameter.

Syntax examples

The following instructions given as examples are to be used within the estimated_params block.

Declaration of a prior for estimated parameter alpha, specified using mean and variance:

     alpha.prior(shape=beta,mean=0.3,variance=0.1^2);

Declaration of a prior for estimated parameter alpha, specified using the cumulative distribution function:

     alpha.prior(shape=beta,interval=(0.2,0.4,.9);

with this syntax the prior is such that 90% of the prior mass lies between 0.2 and 0.4.

Declaration of the subsamples (must preceed the declaration of the priors):

     alpha.subsamples(name1=1950Q3:1957Q4, name2=1958Q1:1983Q2, name3=1983Q3:2011Q2)

Declaration of alpha's prior over different subsamples (must follow the declaration of the subsamples):

     alpha.name1.prior(shape=normal,mode=0.30,stdv=.01);
     alpha.name2.prior(shape=normal,mode=0.33,stdv=.01);
     alpha.name2.prior(shape=normal,mode=0.40,stdv=.01);

-  ⇤ ← Revision 8 as of 2011-09-22 14:25:27 → 
  Size: 2392
  Editor: StéphaneAdjemian
  Comment:
+   ← Revision 9 as of 2011-09-22 14:46:22 → ⇥
  Size: 2556
  Editor: StéphaneAdjemian
  Comment:
-Deletions are marked like this.
+Additions are marked like this.
 Line 10:
-Rely on options names rather than argument position in the command line. This new interface would replace '''estimated_params_bounds''' and '''estimated_params_init'''. The following option names would be necessary:
 * prior_shape
   Specification of the prior distribution. Name of the prior distribution.
 * prior_mean
   Specification of the prior distribution. Mean of the prior distribution. 
 * prior_stdev
   Specification of the prior distribution. Standard deviation of the prior distribution.
 * prior_variance
   Specification of the prior distribution. Variance of the prior distribution.
 * prior_mode
   Specification of the prior distribution. Mode of the prior distribution.
 * prior_interval(lower_bound,higher_bound,percentage)
   Specification of prior distribution from the cumulative distribution function. 
 * lower_location
 * upper_location
 * jscale (Metropolis proposal scaling for a single parameter, little used)
 * init_val (for the optimization routine)
 * boundaries=(lb,ub)
 * this would replace estimated_params_bounds and estimated_params_init
+Rely on options names rather than argument position in the command line. This new interface would replace '''estimated_params_bounds''' and '''estimated_params_init'''. The following option names would be necessary
 * For the declaration of the priors
  * '''shape''', name of the prior distribution.
  * '''mean''', mean of the prior distribution. 
  * '''stdev''', standard deviation of the prior distribution.
  * '''variance''', variance of the prior distribution.
  * '''mode''', mode of the prior distribution.
  * '''interval''', specification of the prior distribution from the cumulative distribution function defining an interval covering an arbitrary percentage of the prior mass. See below for an example. 
  * '''shift''', to shift the domain a distribution. See below for an example (useless if the domain of the distribution is the set of real numbers).
  * '''domain''', specify the domain of distributions with bounded support (uniform and beta).
 * For optimization options
  * '''boundaries''', enforce the optimizer to look for a mode in a specified interval.
  * '''initial''', initial value for the optimizer.
 * For the mcmc options
  * '''jscale''', scale paramemeter for the jumping distribution specific to a parameter.