U.S. Treasury, Office of Financial Research, Tues 9 Apr afternoon (I don’t actually know exactly when or in what room):
Parameterization and Bayesian Modeling
—
Johns Hopkins University, Department of Biostatistics, 4pm Wed 10 Apr, room W2030 School of Public Health:
Little data: How traditional statistical ideas remain relevant in a big-data world
At the end of the day, after all the processing, big data are being used to answer little- data questions such as, Does an observed pattern generalize to the larger population?, or Could it be explained by alternative processes (sometimes called “chance”)? We discuss some recent ideas in the world of “little data” that remain of big importance.
Didn’t have enough notice to take off from work to see you in DC. Thanks for posting the slides!
Thanks for sharing. The Johns Hopkins presentation looks especially interesting. Please update us if a video becomes available to go along with the slides!
Thanks Andrew, they look like good talks. Parameterization in particular, because it’s part of what puts people off Bayesian modelling at an early stage. They return to the comfortable world of t-tests and such, muttering about “dark arts”. The more practical advice on this, the better!
Blue, Red and Purple! What color scheme is that?
Very interesting!
Videos?