CODEWITHSUNDEEP
regexnlpnltkstanford-nlpgate
Viper
Viper

Reputation: 201

Annotator for Relationship Extraction

I have a set of urls in a text file. For each url in that text file, I want to tag the entities and relationships in the text contained in that url.

I am aware of the entity taggers like Stanford NER, NLTK and GATE which can perform the entity tagging. However, I am more interested in relationship extraction.

In order to extract relationships, I am thinking of annotating the text contained in those urls for training purpose. For this, I do not want to do manual annotation. I can write few regex to extract the relationship which I want, however it would be difficult to scale up.

Is there a tool where in I can specify what I want to annotate?

For example:

" Rob is working as the Director of ABC organization. He graduated from XYZ University "

Here, I want to extract the affiliations relationship, so intuitively I would like to annotate words which describe the affiliations like working, graduated.

Edit: By "a set of URLs in the text file", I mean I have about 200 links to certain webpages in that text file, each of the webpage contains some text. I want to analyse (annotate) that text.

Upvotes: 0

Views: 746

Answers (1)

Chester Mc Allister
Chester Mc Allister

Reputation: 437

There is no PR in GATE that that will pair arguments and create instances for you. You must therefore create instances that are relevant to your problem.

You can:

  • write a custom PR
  • or write some JAPE with Java RHS

You can probably split your corpus on a training and a test dataset.

You can use the GATE training course about Relation Extration that contains all you need:

Upvotes: 1

Related Questions