Model evaluation chunk - Githubissues

KIPAC / StatisticalMethods

Course notes and resources for Stanford University graduate course PHYS366: Statistical Methods in Astrophysics

https://kipac.github.io/StatisticalMethods

Other

111 stars 78 forks source link

Model evaluation chunk #138

Closed drphilmarshall closed 7 years ago

drphilmarshall commented 7 years ago

Still thinking about exercises to do in class. Possibilities include:

Play around with values of chi-sq and Ndof to develop some intuition
Derive the E = f Lmax result and explore

PS. This is in response to #99

abmantz commented 7 years ago

You forgot about the Deviance Information Criterion, which is better at accounting for complexity than AIC and BIC. It doesn't look like the notebook flow is really finished at this point, so I'll hold off on other comments.

drphilmarshall commented 7 years ago

No, go ahead: I would happily attempt to teach off what is checked in, so your full feedback would be most welcome! :-)

abmantz commented 7 years ago

Can you clarify whether/when you are planning to jump off into the linked notebooks?

drphilmarshall commented 7 years ago

No, the idea is that it's just there in the background providing the code that made the figures (and computed the numbers).