aurelberra / stopwords Goto Github PK

@aurelberra—thanks for getting this conversation started. Quick update—the words given as CLTK stopwords are not intended to be final results. These are preliminary results collected for a conference paper on ongoing work. I have some updates that I can present soon (and contribute here), likely within the month (sorry, busy October!). Accordingly, I've opened this issue to keep track on my future contribution.

Design shorter, custom stoplists for specific uses? They could be for example:

• Minimal lists (e.g. for search engines): based on frequency only, or including the paradigms of articles and a few verbs?
• Generic lists: separate versions for epic poetry, tragedy, oratory, theology, etc.?

Add most common proper names in Greek and Latin lists? This means compiling NER lists and determining thresholds.

Compare MFW lists with TLG MFW and my current Greek stopwords to make lists for:

• Iliad
• Odyssey
• Iliad and Odyssey

Maybe integrate the new stopwords in the current Greek list.

Bonjour @aurelberra
Si tu as le temps, il me semble qu'un CITATION.CFF ( https://citation-file-format.github.io/cff-initializer-javascript/# ) serait bienvenu :)

Merci pour votre liste excellent! C'est un travail très utile.

En parcourant la liste, j'ai trouvé une erreur mineure: le mot νῷ est inclus dans la liste pour ἐγώ. Bien que cela soit sans aucun doute correct pour ce cas rare, la même forme représente le plus souvent le datif de νοῦς (par exemple dans des phrases comme ἐν νῷ εἶχον) et devrait donc peut-être être supprimée de la Stoplist.