Random Intercept BART - #9
Merged
Merged
Conversation
I modeled the declaration of tausq and tausq burnin after sigmasq
…eady Once again for declaration of statements for tau, I copied those statements for sigma closely.
This was modeled after get gibbs sample for sigsq closely.
Once again, I closely modified from sigsq
This is the main modification to get the random intercept parameters of b_i and tausq. The estimated parameters are likely not consistent with the truth since I have not been able to test the codes in R. However, we have a developed a method that produces consistent estimates by using bart from BayesTree to draw the regression trees and then drawing the sigsqs, tausqs, and bis outside of bart. Hence, I should be able to make the correct changes once I know how to get modifications on bartMachine to work.
Similarly, copied closely from sigsq
Added tausq stuff by modifying from sigsq codes and for grouping variables, copied cov_split_prior.
Updates so that R can recognize the addition of grouping variable and initial values for tau.
Similarly to update R files so that groupings variable can be read into the java files.
kapelner
added a commit
that referenced
this pull request
Jan 13, 2016
Random Intercept BART from Vincent
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This was an attempt to extend the original BART to accommodate for correlation between subjects. I used a random intercept here for simplicity. As I was only able to get the codes to compile successfully in java but failed to run in R, I was not able to test these codes.
The method for drawing the b_i and tausq will likely produce biased estimates from our tests using bart from BayesTree. However, we have a method that works correctly and once we figure out how to correctly modify bartMachine to accommodate for a random intercept, we shall update this code with the correct method.