0
$\begingroup$

I am planning to take up an interesting NLP project. I want to extract 'goal' statements from lengthy reports. For example, the goals can be We would be reducing our carbon footprint by 50% by 2025 or Our company aims to increase the diversity in the work-force in upcoming months. Check below image for example text and highlighted goals.

Example with highlighted goals

How can I go about the process of goal extraction, I would like to get some pointers on possible NLP approaches ?

Improve this question
$\endgroup$
3
  • $\begingroup$ There should be something out there in the NLP literature. Nonetheless, If the problem is about which sentences are 'goals', you could split your text(s) into sentences and then frame it as a Classification task; which sentences are 'goals'. $\endgroup$
    – 20-roso
    Dec 12, 2022 at 17:51
  • $\begingroup$ @20roso, the problem is I do not have annotated data to train the model for classification. $\endgroup$
    – air cooled
    Dec 13, 2022 at 9:28
  • $\begingroup$ In that case, like the answer suggested you can look into a rule-based approach. You can use NLTK or spaCy to check their POS tag (e.g. Pronoun), Dependency Parsing or/and an already trained NER model to make rules based on a sample of sentences. $\endgroup$
    – 20-roso
    Dec 13, 2022 at 15:01

1 Answer 1

Reset to default
2
$\begingroup$

Well, A quick approach to this is using named entity recognition and POS tagging to identify key phrases in the text that are likely to be goals.

For example, you might look for phrases that contain named entities, such as organizations or locations, and verbs that indicate an action or intention, such as "aims to" or "will reduce."

You can learn about that here:

Named Entity Recognition Tagging

Part of Speech Tagging

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.