Exercise

Efficient phrase matching

Sometimes it's more efficient to match exact strings instead of writing patterns describing the individual tokens. This is especially true for finite categories of things – like all countries of the world.

We already have a list of countries, so let's use this as the basis of our information extraction script. A list of string names is available as the variable COUNTRIES. The nlp object and a test doc have already been created and the doc.text has been printed to the shell.

Instructions

100 XP
  • Import the PhraseMatcher and initialize it with the shared vocab as the variable matcher.
  • Add the phrase patterns and call the matcher on the doc.