search

statnet / lolog

Latent Order Logistic (LOLOG) Graph Models
Other
5 stars 1 forks source link

Changes made for Bayesian LOLOG #4

Closed duncan-clark closed 3 years ago

duncan-clark commented 3 years ago

I made these changes to facillitate fully Bayesian fitting of the LOLOG model. Main changes were exposing the dyadUpdate method to R (probably need to fix this so uses R indexing in R ?), and allowing for simulation using full edge orderings, rather than node orderings, and returning change stats when carrying out simulations., this allows for stepping around posterior parameter space.

Code for RMH MCMC stepping around posterior space currently separate.

nfultz commented 3 years ago

Hello, you should probably remove the dotfiles from your PR (.Rdata, .Rhistory).

duncan-clark commented 3 years ago

yes, thank you. To be honest I also did not realise that the LOLOG repo is part of statnet now

ifellows commented 3 years ago

I suggest resubmitting, removing unnecessary files and checking that indentations match the rest of the c++ code.