For anyone who wants to dig a bit deeper: PRML by Bishop [0]. An amazing work as a general introduction to machine learning and Bayesian in general. MacKay's book [1] is a bit more opinionated regarding Bayesian methods. Even if you do not become purely Bayesian, it advances your understanding of to approach data quiet a bit compared to the generic "oh data, lets throw a NN at it".

[0] https://www.microsoft.com/en-us/research/people/cmbishop/prm... [1] http://www.inference.org.uk/mackay/itila/book.html

About 40 years ago I was reading a business mathematics book and ran across a section that, as I recall, started out with a basic probabilities equation and derived the basic Bayesian equation in one step.

Am I remembering this right? If so could someone post it?

By definition: p(A|B) = p(A,B) / p(B) (1)

With the same definition, switching A and B roles, we have p(A, B) = p(B|A) p(A) (2)

Plug (2) into (1) and you have Bayes rule.

Thank you for the explanation.

I would really like to read the full paper for reason #3, but it’s paywalled. The whole article feels like a big teaser with the real content locked away.

The TL;DR of the article provides two links to the preprint, and in fact there is a Springer version accessible without cost: https://link.springer.com/content/pdf/10.3758/s13423-017-131...

Even less friction than Sci-Hub is the Unpaywall browser extension http://unpaywall.org/products/extension which finds links to unpaywalled content matching paywalled journal articles.

scihub!