I recently wrote an article on our company blog where I explained how we evaluate Machine Learning models at RecordPoint – if you haven’t read it yet, I’d recommend you do so before proceeding.

Since its publishing, I received a number of comments asking me to clarify the concepts presented with concrete examples.

The original article is light on code by design as the audience for the company blog isn’t always that technical. Therefore, in this post I’ll provide a simple example that will hopefully clear things up – in particular why accuracy isn’t enough.