Wishlist: 2d positions - Githubissues

kbenoit / wordshoal

quanteda implementation of the Lauderdale and Herzog (2016) "Wordshoal" model

13 stars 2 forks source link

Wishlist: 2d positions #1

Open kbenoit opened 6 years ago

kbenoit commented 6 years ago

Issue by methodds Friday Jan 13, 2017 at 12:01 GMT Originally opened as https://github.com/kbenoit/quanteda/issues/452

Hi, it's really nice to see that the 1d implementation of wordshoal is already available in quanteda. Can you please also provide support for the 2d model as described in Lauderdale and Herzog 2016?

kbenoit commented 6 years ago

Comment by kbenoit Friday Jan 13, 2017 at 12:09 GMT

Up to @lauderdale ...

kbenoit commented 6 years ago

Comment by methodds Tuesday Mar 14, 2017 at 14:45 GMT

As another suggestion for improvements, it would be nice if wordshoal models could include feature names and effects marginal effects. I cannot find this in the documentation:

tol: log-posterior tolerance used in fitting
dir: global identification of the dimension
theta: estimated document positions-
beta: debate marginal effects
alpha: estimated document fixed effects
psi: estimated document debate-level positions
groups: document groups
authors: document authors
ll: log likelihood at convergence
se.theta: standard errors for theta-hats
data: corpus to which the model was fit

kbenoit commented 6 years ago

Comment by kbenoit Tuesday Mar 14, 2017 at 15:50 GMT

@methodds this sounds like a separate issue - but I was not entirely clear on what you meant. Perhaps create a new issue and clarify?

kbenoit commented 6 years ago

Comment by methodds Tuesday Mar 14, 2017 at 15:53 GMT

I just meant the wordshoal "equivalents" of @features, @beta, @psi in textmodel_wordfish(). You still want me to open a new issue?

kbenoit commented 6 years ago

Comment by kbenoit Tuesday Mar 14, 2017 at 15:54 GMT

Yes since it's not about implementing a two-dimensional version and its resolution is independent of that issue.

kbenoit commented 6 years ago

Comment by methodds Tuesday Mar 14, 2017 at 16:02 GMT

okay, I opened #598

MartinsRodrigo commented 4 years ago

Hi! Just supporting the idea about 2d implementation.