I have a clarification. I have to create a classification model for certain set of documents. We are supposed to flag it anamoly or not based on certain terms in the document. My question is the terms maybe influenced by multiple parameters. Let us say the properties of the document will give different font sizes even though it refers to same term distributed throughout the document.

So let us assume I have a term1 occuring on the first page of the document, I will extract the properties of the font. Then again, I will look for same token and if the properties of the term1 does not match in the other pages, I will flag it as anamoly.

How do we represent them in form of feature vectors for classification model?

I need to incorporate the properties of the terms as one of the input vectors in the document.

Any help would be appreciated.

  • 1
    $\begingroup$ Welcome to DataScienceSE. it looks like the specification of the task is not clear. What kind of document is it about? Is the task about all the terms in the document? Why does the formatting matter? Technically it's possible to add features which represent format properties, but I doubt the system would work as expected. $\endgroup$ – Erwan Feb 27 at 18:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.