bst-mug / acres

Acronym expansion module based on word embeddings and filtering rules
Apache License 2.0
1 stars 2 forks source link

Lowercasing #71

Open michelole opened 5 years ago

michelole commented 5 years ago

Lowercasing might make models denser and solve some issues, e.g. ACS = akutes_Koronarsyndrom.

We might need to relax filtering rules for a case-insensitive match as well.

michelole commented 5 years ago

We tried that informally and had worse results, downgrading.

michelole commented 5 years ago

Another idea would be to lowercase everything that it's not an acronym and accepting matches on lowercased versions of expansions.