tvall / emoxicon Goto Github PK

View Code? Open in Web Editor NEW

0.0 0.0 0.0 142.47 MB

Repository for the Emoxicon R package

R 100.00%

emotion-detection item-response-theory rasch-model text-analysis

emoxicon's People

Contributors

Watchers

emoxicon's Issues

warnings and errors

Should we save all warnings and errors for the individual models, or just let them print to the screen? Definitely we should save errors b/c otherwise if you run 500 you have to track down which one broke.

Staged installation

Something is causing this package to fail staged installation. I can't find anything from my code that would cause this; it could be an issue with package dependencies? For now, we will opt out of staged installation.

Cleaning trolls data

Need to have the trolls data have the accounts with less than 30 removed from the dataset (maybe? or leave it?). Right now they're in so I can test the code.
The cleaning file for the trolls data links to my dropbox, remove

Limiting everything to 30 tweets at minimum

Right now, the code lets you run the parent rasch model on all the tweets. Then, when it runs the individual models it will cut out tweets from accounts with less than 30 posts. So the parent model contains more data than the individual models, collectively.

Should we allow this? Or should we move the cutting to the beginning (assuming that groups is specified).

Recommend Projects