Given the "house built on a hill" example, where the model may incorrectly predict that being built on a hill might be worth more than the square footage--isn't that the type of insight we are looking to gain from this technology? For the computer to identify correlations among the data that we currently discount or are otherwise oblivious to? I can see a correlation, for example, of houses built on hills to provide extra value due to vistas, flood deterrent, etc. (I realize it was just an example, but the point seems to be to have the machines learn new correlations for us, right?)
Evaluation metrics is a great way to look at validity. There was a time where vanity metrics was what was presented. ML adds another layer, and better nuance for decisions because of the nuance in the model results.
Get up to speed with emerging analytics technologies including Natural Language Processing, Edge Analytics, Machine Learning, Real-time Analytics, and Augmented Analytics. These expert-led sessions are for analytics leaders, professionals and business users.
Get started with tomorrow's analytics technology. Sign up today.