Here we discuss the secret metric mix used by air controllers to decide which flights get cancelled, as the administration has accepted to reveal how the "cancellator" algorithm works.
It is interesting that they call it "predictive modeling" (a branch of statistics) when actually it is "mathematical optimization" (a branch of operations research). They use Monte Carlo simulations to identify the best scenarios. Note that both predictive modeling and mathematical optimization (applied to real data) are sub-domains of data science. It proves that the same results can be obtained via very different approaches: one based on statistical modeling (predictive modeling; hierachical Bayesian models would work well in this context), and one approach where probabilistic reasoning seems to be a-priori absent but actually implicit (mathematical optimization).
Anyway, check the metrix mix at http://goo.gl/4CyEBn
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|