example situation, possibly for illustration and problem generation

email dated 2013-01-17

Audit the texts for this word and footnote it, at least.

Email response from me 2013-05-06

• the scale dependency of the Bayes Factor (i.e. the evidence) when using the line angle instead of the slope as a parameter if the data points have units.
• the issue of hitting the lower limits of the uncertainties in both directions with the posterior probability distribution, because the model does not know where to put the scatter. I showed you a corner plot with this effect. You suggested to give not only upper limits but also sensable lower limits to the uncertainties.
• the question of Markus about how one can put into a model the difference between all data points lying on one side of the line or equally distributed on both sides of the line. In our case we couldn't tell our model that it is an important finding that all redshifts lie above zero.

Try this; is it better? If so change.

I have various "wrong" equations. Include one that involves multiplying together posteriors, as is done, wrongly, in this paper: http://arxiv.org/abs/1604.02433

I believe there is a typo in

    When the uncertainties are Gaussian and their variances $\sigma_{yi}^2$

I also found this problem to be poorly worded

The statement

setting all $\sigma_{yi}^2=S$, that is, ignoring the uncertainties and replacing them all with the same value $S$

is not self-consistent.
I believe it should read as

\begin{problem}\label{prob:chi2} 
 Re-do the fit of \problemname~\ref{prob:easy} but setting all 
 $\sigma_{yi}=S$, that is, ignoring the uncertainties and replacing 
 them all with the same value $\sqrt(S)$.  What uncertainty variance $S$ would 
 make $\chi^2 = N-2$?  Relevant plots are shown in 
 \figurename~\ref{fig:chi2}.  How does it compare to the mean and 
 median variance of the uncertainty $\allsigmay$? 
 \end{problem} 

...and add them into pgm.tex with text around them.

There is straightline code.

There is modelcomplexity code.

should be on README

should be on PGM chapter (is)

should be on probability.tex if/when he adds a PGM to that chapter

so that I can import the chapters preserving the (informative) history.

write some notes, including various Luptonisms.

just a tiny style point

• assigned to @dfm, who will do this via pull request
• must acknowledge Geert, who first suggested it.

I am pissed, because I thought the import from SVN was supposed to preserve it; and it did at the foo stage; the git fetch and git merge seemed to eradicate it...?

@HWRix says that any highly cited paper should be IN A REFEREED JOURNAL.

given changes made in section 7

be sure to set a license!

make sure scope is correct and names are correct

should mention likelihood and posterior

In email from Dan to Hogg dated 2012-03-30. I need to print out, digest, edit, acknowledge, commit, and push

I can't figure out how to get to the main point quickly!

Also, we should give a real-world example; maybe the Astrometry.net one at least...

I now don't like Data Analysis Recipes. Other options:

• Data Analysis
• Data Analysis for natural sciences and engineering
• The Joy of Data Analysis
• ... add suggestions!