AD Research Wiki:

Our and other tools for Named-Entity Recognition and Disambiguation

Contents

  1. TagMe
  2. Ambiverse

TagMe

Register here to get access to the TagMe API. Once logged in, click 'show' on the left panel to get your personal API Token.

You can then query the API with a request as follows, replacing <TOKEN> with your token and <TEXT> with the text to annotate:

https://tagme.d4science.org/tagme/tag?lang=en&gcube-token=<TOKEN>&text=<TEXT>

The response is a JSON object. See this example (using Matthias' token).

Each annotation has a certainty level rho (between 0 and 1). According to the documentation, reasonable threshold values are between 0.1 and 0.3.

Check the API documentation for help.

tagme-python is a simple wrapper to query the API and parse the responses.

The original source code of TagMe is available on Github.

Ambiverse

TODO

AD Research Wiki: Projects/NER+NED (last edited 2020-08-28 14:01:36 by Matthias Hertel)