How do I output only extractive Key phrases
This is a really great fine tuned model , I would like to know how to separate out only the extractive keywords from the abstractive ones ? thanks
Hi!
That's easy to do. I recommend that you split the text into a word list and remove all punctuation marks.
Then use the Python in operator to see if the keyphrase is in that list. If the key phrase is in the list, it is extractive, otherwise it is abstractive.
I hope this helps!
Hi!,
Thanks for the reply , I was curious about if there is a way i can tweak the pipeline maybe by giving an argument or so to extract only the extractive/abstractive key phrases , sometimes splitting the words and checking we might miss out certain -ngrams keyphrases ..
Thanks!
Hi!
This is indeed a fascinating question! I forgot to mention that you can generate a list of n-grams so you don't miss any n-gram keyphrases. More information here: https://thinkingneuron.com/how-to-generate-n-grams-in-python/.
What I also want to add is that this is a keyphrase generation model. The purpose of these types of models is to generate keyphrases that (not) appear in the text. So you can discuss about this feature. If your focus is on extractive keyphrases, you should use an extractive keyphrase model instead. Keyphrase extraction models perform a lot better than keyphrase generation models.
I'm very intrigued by your question, so I'll try to come up with something. If you have anything to add let me know :)
I hope this helps!