PDA

View Full Version : The terminology of Cosmologists - priors?



tashirosgt
2019-Aug-20, 05:13 PM
In the paper about Cosmological parameters https://arxiv.org/pdf/1709.10504.pdf the term "prior" is used but the paper gives no specifics about prior distributions in the sense of Bayesian statistics. Does the term "prior" have a technical meaning in Cosmology that is distinct from its use in statistics?

This question arises because someone in another forum (physicsforums.com) asked for an interpretation of charts in that paper. The question was asked in the section on probability and statistics. I'm wondering if the question belongs in that section.

Shaula
2019-Aug-20, 06:17 PM
Its hard to say for sure as the reference article is covered by a paywall (or my Google Fu is weak) but I'd expect that the distribution applied is in the paper by Yu. That paper just appears to give the means. I'm not aware of any special meaning of priors in cosmological modelling otherwise.

StupendousMan
2019-Aug-20, 11:11 PM
Does the term "prior" have a technical meaning in Cosmology that is distinct from its use in statistics?



No, it means the same thing.

If the link to the PDF in the OP is blocked, try this: https://arxiv.org/abs/1709.10504

tashirosgt
2019-Aug-21, 02:51 PM
No, it means the same thing.

If the link to the PDF in the OP is blocked, try this: https://arxiv.org/abs/1709.10504

Is the usage of priors so standardized that cosmology papers don't have to explain what priors they are using? That paper seems to refer to a contant value as a "prior". How are they getting a Bayesian prior from one constant value?

StupendousMan
2019-Aug-21, 08:45 PM
I've just taken the time to scan the paper. I agree with tashirosgt: in this instance, the authors appear to use the "prior" to refer to the fixed value of some parameter in a simulation. That's not a very good practice.