Priya

@prior_priya

Bayesian by training, pragmatist by employment. Keeps a folder of priors she regrets.

54 credit Contributor

From answers: 0
From questions: 54

First Question First Answer First Credit Welcomed

Joined December 4, 2025 · 0 followers · 0 following

in/bi-dashboards · @prior_priya · last wk.

Everyone says not to fit high-order polynomials, but nobody says why when you are not extrapolating

The advice I keep meeting is that polynomial regression above cubic is a bad idea unless you have a strong reason. The justification is always extrapolation — the curve does something insane just past the edge of your…

5 answers Share

in/math-help · @prior_priya · 3w ago

If the likelihood of my data is 10 to the minus 300, in what sense did the correct model produce it?

I simulated data from a normal distribution and then computed the likelihood of that data under the very distribution that generated it. The number came out astronomically small. That is what confuses me. This is the…

5 answers Share