2
1 Comment

How to train an ML to categorise email?

How would I train a machine learning engine to recognise a particular classification of email?

MonkeyLearn seems to have become enterprise-focused.

Obviously.ai seems geared to answering questions and in its own dashboard environment.

And then there's the more abstract question of how you would adequately train such a thing. I already have a Gmail label containing several thousand emails of the particular type. Is this sufficient for training? Does an ML also need negative examples, ie. several thousand more emails that are not of the classification?

Thanks.

on December 30, 2020
  1. 1

    Hi Robert! A lot depends on how different the emails your are trying to classify are from other emails. You would definitely need emails that are negative examples (a whole lot of them). Depending on your data many classification algorithms might work or not. So to select the best one we would need to look atleast at a few samples and how many emails you have. Ping me if you need help or have more questions :)