documentation of bootstrapping in dic.fit()
Closed this issue · 4 comments
Since we're adding asymptotic distributions for some models, we need to clarify more clearly how you specify which SE to calculate and how many bootstraps to run.
can you take care of this or do you need me to?
i need clarification from you on this, but i can take care of it. is this correct:
dist -- methods available for SE -- fitting methods available
L -- asymptotic, bootstrap -- optim, MCMC
W -- asymptotic, bootstrap -- optim, MCMC
G -- bootstrap -- optim, MCMC
E -- NA -- MCMC
If the above is correct, then I'm just thinking that we could change the logic in some of the error-checking statements in dic.fit() to indicate that one could choose to specify n.boots=0 for L or W.
Nicholas G. Reich, Ph.D.
Research Assistant Professor
Division of Biostatistics and Epidemiology
School of Public Health and Health Sciences
University of Massachusetts, Amherst
413.545.4534
nick@schoolph.umass.edu
http://people.umass.edu/nick/
on twitter: @reichlab
On Tuesday, July 9, 2013 at 9:36 PM, Andrew Azman wrote:
can you take care of this or do you need me to?
—
Reply to this email directly or view it on GitHub (#13 (comment)).
Looks good to me. I hope it is clear from the table that they will still get credible intervals from the MCMC.
Yeah we may want to make the input options a bit more verbose so that n.boots is not both the flag for bootstrapping and the number of bootstraps.
On Jul 9, 2013, at 9:40 PM, Nicholas G Reich wrote:
i need clarification from you on this, but i can take care of it. is this correct:
dist -- methods available for SE -- fitting methods available
L -- asymptotic, bootstrap -- optim, MCMC
W -- asymptotic, bootstrap -- optim, MCMC
G -- bootstrap -- optim, MCMC
E -- NA -- MCMCIf the above is correct, then I'm just thinking that we could change the logic in some of the error-checking statements in dic.fit() to indicate that one could choose to specify n.boots=0 for L or W.
Nicholas G. Reich, Ph.D.
Research Assistant Professor
Division of Biostatistics and Epidemiology
School of Public Health and Health Sciences
University of Massachusetts, Amherst413.545.4534
nick@schoolph.umass.edu
http://people.umass.edu/nick/
on twitter: @reichlabOn Tuesday, July 9, 2013 at 9:36 PM, Andrew Azman wrote:
can you take care of this or do you need me to?
—
Reply to this email directly or view it on GitHub (#13 (comment)).
—
Reply to this email directly or view it on GitHub.
Correct, and good clarification. I'll work on this tomorrow hopefully.
Nicholas G. Reich, Ph.D.
Research Assistant Professor
Division of Biostatistics and Epidemiology
School of Public Health and Health Sciences
University of Massachusetts, Amherst
413.545.4534
nick@schoolph.umass.edu
http://people.umass.edu/nick/
on twitter: @reichlab
On Tuesday, July 9, 2013 at 9:52 PM, Andrew Azman wrote:
Looks good to me. I hope it is clear from the table that they will still get credible intervals from the MCMC.
Yeah we may want to make the input options a bit more verbose so that n.boots is not both the flag for bootstrapping and the number of bootstraps.
On Jul 9, 2013, at 9:40 PM, Nicholas G Reich wrote:i need clarification from you on this, but i can take care of it. is this correct:
dist -- methods available for SE -- fitting methods available
L -- asymptotic, bootstrap -- optim, MCMC
W -- asymptotic, bootstrap -- optim, MCMC
G -- bootstrap -- optim, MCMC
E -- NA -- MCMCIf the above is correct, then I'm just thinking that we could change the logic in some of the error-checking statements in dic.fit() to indicate that one could choose to specify n.boots=0 for L or W.
Nicholas G. Reich, Ph.D.
Research Assistant Professor
Division of Biostatistics and Epidemiology
School of Public Health and Health Sciences
University of Massachusetts, Amherst413.545.4534
nick@schoolph.umass.edu (mailto:nick@schoolph.umass.edu)
http://people.umass.edu/nick/
on twitter: @reichlabOn Tuesday, July 9, 2013 at 9:36 PM, Andrew Azman wrote:
can you take care of this or do you need me to?
—
Reply to this email directly or view it on GitHub (#13 (comment)).—
Reply to this email directly or view it on GitHub.—
Reply to this email directly or view it on GitHub (#13 (comment)).