Enlarge / Turn the lens on ourselves, so to speak.

There is a moment in every foray into new technological territory that you realize that you may have been engaged in a Sisyphean task. Given the myriad of options that are available to you to take on the project, research your options, read the documentation, and get down to work – only to find that it’s actually just define the problem can be more work than finding the actual solution.

Reader, here I found myself two weeks into this machine learning adventure. I became familiar with the data, the tools, and the known approaches to solving problems with this type of data, and tried several approaches to solving a seemingly simple machine learning problem: could we based on past performance? predict whether a particular Ars headline will win in an A / B test?

Things weren’t going very well. Indeed, like me …

Source link

Leave a Reply