New answers tagged

1

It's true that it's a bit of a complex process but it's worth understanding it in order to get the best out of the model. "Feature" and "attribute" (and probably observation but I'm not 100% sure) are the same thing. The features are the ones directly used by the model (as opposed to the raw input data). For every input word a vector of ...


2

One can find sequence labelling libraries by searching for the term conditional random fields, the state of the art method. Probably one could also find libraries and tutorial by searching the term Named Entity Recognition, which is certainly the most standard NLP application of sequence labelling. Here are a few libraries that I know of: CRF++ crfsuite (...


Top 50 recent answers are included