I have a large number of different XML documents with associated style sheets (SEC financial filings). Are there any natural language processing (or other) tools that I can use to try to automatically extract section headings?

  • $\begingroup$ Are the section headings tags in the xml, or are you trying to derive them via natural language processing methods? $\endgroup$
    – Kyle.
    Commented Nov 13, 2015 at 19:39
  • $\begingroup$ There will inevitably be some kind of tags associated with the headings, but its not standard between documents. Here's one example (which happens to be html) sec.gov/Archives/edgar/data/320340/000143774915020747/… $\endgroup$ Commented Nov 13, 2015 at 19:54
  • $\begingroup$ NLP is not a tool to extract headings. Extract the text and use tags as features. And html is not the same as xml - it will be different tools for extraction. $\endgroup$
    – paparazzo
    Commented Nov 13, 2015 at 20:25

1 Answer 1


It seems you are lucky and the section headings are XML-tagged. So you can use XSLT transformations or Python lxml.etree to get rid of them or to extract them.


Your Answer

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

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