“When would you use a Microsoft Syntex document processing model classifier rather than a Trainable Classifier to apply a retention label?”
I have been asked this question several times and I stumbled over the answer long enough to realize I needed to be able to articulate the answer in a more coherent way. I also want to extend the answer to apply to more than just retention labels… there are differences across the board in some key areas I see the two types of classifiers being used.
This post is not about how to configure these 2 types of classifiers, Microsoft has done a great job at doing that already:
Although both a Microsoft Syntex model classifier and a Trainable classifier can certainly identify content, the capabilities and what you can do with the content once it’s classified vary greatly.
I decided to build a table to highlight some of the similarities and differences. You may find this helpful when trying to decide which is the appropriate option for your scenario.
Link: Syntex Classifier V Trainable Classifier
What else would you include? I’m always looking for feedback! 🙂
Thanks for reading.
Great post Joanne! I’ve been meaning for a couple of years now to try to run a practical side by side test to see whether Trainable Classifiers or Microsoft Syntex would label with more accuracy (guess it entirely depends on the way the Syntex model is defined), but assume that Syntex would usually prove the more reliable approach.
Hey Rob! Thanks! Interesting question. I’m not sure which one would be more accurate… so many factors at play that wouldn’t make it a neat answer… file types, formats and variations, model type used, etc. It would be a good factor decision tree to see.