Email NLP Entity Parsing
Hi there!
Currently, I have a keyword model called "Action" that picks up: "meet with", "talk to", and "visit with."
Also, I have the standard person entity extraction model to pick up names like "Joe", "Steve", and "Ryan."
The common format for emails that I receive are:
"Ryan would like to meet with Steve"
"Joe would like to talk to John"
"Rick would like to visit Roger"
The NLP extraction will return:
"Ryan" - #Person
"meet with" = "Action"
"Steve" - #Person
My goal is for:
"Ryan" = #Person
"Ryan" = #Visitor
"meet with" = #Action
"Steve" = #Person
"Steve" = #Host
So logically, the #Person before #Action should be the #Visitor and the #Person after #Action should be the #Host.
What is the best way to implement this in the Text Analyzer/NLP Model?
Thanks,
Ryan
***Edited by Moderator: Lochan to update platform capability tags***
Pegasystems Inc.
IN
Hello Team,
So collecting the names and places and few more things are done by entity extraction and for the actions they use classification for the pattern so you should use the entity extraction for the patterns. so upload the training data in the decision data and add this decision data as entity model to the text analyzer.
Regards
Prasanth