Make model_averaging.ipynb runnable, fix formula typo

[MarcoGorelli]

Couple of things:

• original notebook accessed a local csv file, so I've replaced that with one from pymc/resources
• pretty sure the formula reported had a typo, as we're maximising the weights rather than the number of points (the original paper at least has w instead of N)

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[MarcoGorelli]

The only thing I changed here was to maximise over w rather than over N

Reply via ReviewNB

[codecov]

Codecov Report

Merging #4271 (f6db586) into master (edbafaa) will increase coverage by 0.02%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #4271      +/-   ##
==========================================
+ Coverage   87.44%   87.47%   +0.02%     
==========================================
  Files          88       88              
  Lines       14342    14342              
==========================================
+ Hits        12542    12545       +3     
+ Misses       1800     1797       -3     

Impacted Files	Coverage Δ
pymc3/parallel_sampling.py	86.79% <0.00%> (+0.94%)	⬆️

[michaelosthege]

LGTM

[MarcoGorelli]

Thanks! Time to cross off exercise 7, chapter 5 from Osvaldo's book:

Read and run the model averaging example from PyMC3's documentation at
https:/​/docs.​pymc.​io/notebooks/​model_​averaging.​html .