2

I am currently working on a clinical NER tagging application (CliNER - Clinical Named Entity Recognition system) and I am using a Bi-Directional LSTM built with Keras / Tensorflow backend for this task.

I want to know whether there is any method for adding a condition to the output.

For E.X. if the previous output was B-Medicine my next should definitely
I-Medicine or Other and not some random tag like B-Body-Part.

Is there a way to accomplish this?

petezurich
  • 9,280
  • 9
  • 43
  • 57
Saul_goodman
  • 141
  • 12
  • What kind of conditions do you have in mind? Like give example. – Daniel Jun 20 '17 at 16:16
  • what do you mean by "previous output"? Is this a seq-2-seq model? If yes, the answer is yes, but you need to be more clear. – Mehdi Jun 21 '17 at 11:21

0 Answers0