IBM Watson™ Discovery Service Ideas

We've moved...

You'll be redirected shortly, we've moved to our new idea portal: https://ibm-watson.ideas.aha.io

Document splitting based on machine learning

Currently in order to split documents we need to painstakingly identify fonts that might be used as headings across all the different documents that we are ingesting.  If the documents have a large variety of fonts and formatting this job becomes very laborious, and it is likely that a heading font in one document would be normal text in another document.

 

Instead WDS could use the same method of identifying what are headings and where to split documents that humans do:

- Bigger fonts with shorter text (compared to the rest of the doc) are probably headings.

- Lower-level headings appear under higher-level headings in the document flow, and have smaller fonts.

- Longer documents should be split up and shorter ones not split up.

 

This does not just effect PDF.  People use Word and HTML without using heading styles / tags as well.

 

Enabling document splitting would then just be a tick box, maybe accompanied by a rough indication of desired document size.

  • JAYSEN OLLERENSHAW
  • Mar 29 2018
  • Shipped
Why is it useful?
Who would benefit from this IDEA? Wow! Watson just KNEW how to split up my documents!
How should it work?
Idea Priority
Priority Justification
Customer Name
Submitting Organization
Submitter Tags
  • Attach files
  • Admin
    Phil Anderson commented
    03 May 17:14

    This is now possible with Smart Document Understanding