Text mining and content categorization

Text Mining

Posts: 46

Text Mining



I am currently building a datatable in base sas and using an index function to flag certain company names embedded in a paragraph of text in a column. If the company name exists I will flag them with a one. When I've looked into the paragraphs in more detail this simple approach doesn't work. Take this example below;


"John Smith advised Coco-cola on its merger with Pepsi". I'm searching on both Coca-cola and Pepsi but only want to flag Coca-cola in this example as John Smith "advised" them. I don't want both Coco-cola and Pepsi flagged with a "1". I understand that I can write code that takes words after certain anchor words such as "advised", "represented" which does work. What happens if one record simply lists all companies that they have advised without using an anchor words to identify them? Is there any tools out there that can do this automatically by AI?




Frequent Contributor
Posts: 80

Re: Text Mining

Have you looked into the FINDW and SCAN functions?

Ask a Question
Discussion stats
  • 1 reply
  • 2 in conversation