Upcoming Courses

On Wednesday, December 10 I will be offering my comprehensive introduction to regression modeling at a steep discount in an effort to raise funds for World Central Kitchen and United Farm Workers.

All you have to do to attend is make a donation of 50 USD to either organization and then send me a screen shot of the confirmation. Details about the course and registration process can be found on my website, https://betanalpha.github.io/courses/.

I greatly appreciating any sharing of the course details to your colleagues and communities!

Consulting and Training

Are you, or is someone you know, struggling to develop, implement, or scale a Bayesian analysis compatible with your domain expertise? If so then I can help.

I can be reached at inquiries@symplectomorphic.com for any questions about potential consulting engagements and training courses.

Probabilistic Modeling Discord

I set up a Discord server dedicated to discussion about (narratively) generative modeling of all kinds. For directions on how to join see https://www.patreon.com/posts/generative-88674175. Come join the conversation!

Support Me on Patreon

If my work had benefited your own and you have the resources then consider supporting me on Patreon, https://www.patreon.com/c/betanalpha. Amongst other bonuses, over the past few months my supporters have enjoyed sneak previews of the material for the upcoming courses.

Recent Rants

Odds ratios are strictly less interpretable, and hence useful, than probabilities when considering a homogeneous behavior. That said odds ratios do have some utility when considering _heterogeneous_ behaviors.

In a logistic regression model linear changes to the inputs correspond to hyperbolic rotations of the outputs, which are far from intuitive. I mean they're literally the same math that goes into adding velocities in special relativity. On the other hand linear changes to the inputs of the logistic function correspond to _proportional_ changes to the odd ratios. While not as clean as linear changes, most people have much more intuition about proportional changes which can make odds ratios useful for tasks like prior modeling.

Ultimately odds ratios aren't perfect, but they're definitely better than assuming linear probabilities. Because we all know that linear probabilities are rarely adequately assumptions. Especially when the case probabilities are small. No matter how much you want linear behaviors. Right?

Anyways for much more see Section 2.2 of https://betanalpha.github.io/assets/case_studies/general_taylor_models.html#22_Interval_Constraints (and the appendix if you want to learn how to impress your friends by dropping "rapidity" in casual conversation).