Are you measuring the process or the inputs? Because in your post the inputs are what you criticized (the prior and the data), not the process itself. Every ma

Are you measuring the process or the inputs? Because in your post the inputs are what you criticized (the prior and the data), not the process itself. Every machine learning algorithm takes these inputs. So I was just saying that it seems unfair to penalize Bayesian statistics for this and not any other method.