meyersbs / uncertainty

A Python implementation of the uncertainty classifier, based on the work of Veronika Vincze.
MIT License
16 stars 4 forks source link

Classifier never outputs 'U' #6

Closed lucas0 closed 3 years ago

lucas0 commented 3 years ago

Used more than 1500 texts, averaging 400 words each and not a single U word was output by classifier. Seems that the classifier is not working, if it ever did.

image

meyersbs commented 3 years ago

It looks like you're using the code in multilabel mode; either you passed the -m flag on the command line or are not passing binary=True to Classifer() in the code. In multilabel mode, individual words will not be labeled with a generic U, but instead with a C for certain or one of the four types of uncertainty:

I see N and E a few times in your output, which is expected in multilabel. If you were to run the same command without the -m flag (or with binary=True), then it would only output C or U.

See the wiki for more information.