Skip to content

Instantly share code, notes, and snippets.

@pwl

pwl/probabilistic_pca.ipynb

Created March 3, 2017 12:07
Show Gist options
  • Save pwl/2f3c3e240b477eac9a37b06791b2a659 to your computer and use it in GitHub Desktop.
Save pwl/2f3c3e240b477eac9a37b06791b2a659 to your computer and use it in GitHub Desktop.
Download ZIP
Probabilistic PCA with ARD
Display the source blob
Display the rendered blob
Raw
probabilistic_pca.ipynb
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
@pswpswpsw
Copy link

pswpswpsw commented Jul 1, 2018 •
edited
Loading

Very nice notebook on using Edward for PPCA.

However, the statements about original working version of PPCA is not very true.

  • First, the original working PPCA [1] does not involve prior on weight, so it treats weights as deterministic variable.
  • Second, even for the original PPCA, there are control over the variance, unlike the N(0,1) in your script.

Ref:
[1]PPCA original paper: https://www.microsoft.com/en-us/research/publication/probabilistic-principal-component-analysis/

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment