
Thank you for a very informative and straight to the point article with a real world use for machine learning.
I've lost count of the number of times I've seen "AI" bandied about for things like toasters (ok, maybe an exaggeration)
I'm already conjuring the data sets for this in my head as it can apply to my line of work though the data cleaning and selection is going to be a pain as if you teach it wrong then it's always going to be wrong. It would be interesting to try and calculate error rates based on the learning input to know roughly what you would be striving for though I would assume you would want 0 errors. On the other hand if you get the input based on your current data and outcome then it would be as good as what you have already. Lots to think about.