polyglot icon indicating copy to clipboard operation
polyglot copied to clipboard

Inaccurate POS tag

Open devikasondhi opened this issue 5 years ago • 1 comments

Hello,

Referring to the text sample We will meet at eight o'clock on Thursday morning. in POS.ipynb notebook, shouldn't the token o'clock be tagged as an adverb instead of a Noun, as can be verified with any standard English dictionary?

devikasondhi avatar Oct 07 '20 15:10 devikasondhi

Probably yes! BUT: POS Taggers are either probabilistic models or rule-based methods and are therefore NEVER 100% accurate. Even modern models using contextualized word embeddings have some 5 - 10 % error rate on this Task :)

floschne avatar Jan 22 '21 17:01 floschne