1

Suppose I have a corpus of sentences, which I then use a part of speech tagger to annotate.

What algorithms or techniques can I then input the tagged data into? Are the tagged words somehow useful just by themselves?

1
  • 1
    If you have a tagged corpus, you can parse the sentences to see how they work. If you have an untagged corpus, you have to tag it before you parse it, or else use a tagger addon to the parser. Parsing rules are written in terms of part of speech, and that's what the tagger determines (mostly -- even the best make weird mistakes).
    – jlawler
    Sep 8 '15 at 18:04
1

The first use to come to mind is syntactic parsing. If you have a rule based grammar where your symbols are POS tags, then you could generate the tree representation of the sentence. Or, if you have a statistical tree parser, then it can use the POS tags as features for determining the tree/parse structure.

In my own experience in NLP in both industry and academic research, I've primarily used POS tags for machine learning features.

They often are valuable features in Named Entity Recognition tasks and other ontology labeling tasks.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.