Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Specifying the number of test cases for wi task #62

Open
sweetmals opened this issue Aug 29, 2021 · 1 comment
Open

Specifying the number of test cases for wi task #62

sweetmals opened this issue Aug 29, 2021 · 1 comment
Labels

Comments

@sweetmals
Copy link

Hi,
Is there a way to set the number of questions for the word intrusion task similar to topic intrusion task? At the moment, for instance, if I am assessing a model with 40 topics, by default the wi test has 40 cases.
I would prefer to be able to set the number of questions for the word intrusion task as well. Chang et al.'s (2008) experiment, they set 10 cases in each task. This is particularly helpful when assessing multiple models in which the number of topics are quite high and the human resources are limited.
Thanks!

@chainsawriot
Copy link
Collaborator

@sweetmals Thanks for the suggestion. It is possible to implement this. But the problem is, there is a crucial difference between Chang et al (2008)'s offering of only 10 cases to crowdcoders and only offering cases less than k in oolong. Chang et al.'s crowdcoding approach will ultimately have complete coverage of all k topics with at least 8 codings from their crowd. However, if we allow wi with "less than k" cases, one can't have complete coverage. I have a hesitancy to implement this in oolong because it gives users a false sense of validity. (let's say a user has a topic model with k = 100 and then this user implements a wi with only 1 case. And then this user reports in his or her paper that the model has been validated with oolong.) If it is to be implemented, there will be a warning message in every step, to say the least.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

2 participants