Using metatagger to scan content

bikerider

Has anyone used metatagger to scan content submitted via the web for inappropriate content, such as profanity and political correctness? The requirements would be:

Once I run a batch of files (pdf, html, doc) through a recognizer, how can I then push a notification to the user that there is content with suspect text?

Migrateduser

This is a great idea!

(and one I will use when talking to customers, I'll make sure to give you credit..)

Off the top of my head, there are a few things you need to do.

1) Analyze existing content for inapppropriate terms. After creating the list of terms as a MT vocabulary, you'll probably want to batch process all of your existing content. I would write a script (Perl would probably do well here) that runs iwgenmetadata and identifies content that contains terms in the vocabulary. You could then either create a filelist of content with inappropriate terms or perhaps even automatically create review workflows.

2) For ongoing content submissions, you'll want to run iwgenmetadata as part of the approval workflow (<external task>) and automatically reject content that contains a term in the vocabulary.

This same concept could be applied in many different ways. For example, you could use this technique as a link checking device.... have MetaTagger extract URLs from content and have a script verify the URLs actually exist.

--
Tom Wentworth
Sales Engineer
Interwoven