Visualizing PCA's Information Loss

  • 900

In this screencast, we look at how much information is lost during a PCA transformation by applying PCA to an image: an interesting approach to image compression.


  • How PCA can be used for image compression (though it shouldn't be!)
  • How the original data is distorted by projecting down to lower dimensions. 
  • How to load images in IPython and convert them to design matrices. 

We Also Recommend