2

I'm running some basic R and there are a lot of great tutorials about PCA but they always stop once they've obtained a biplot.

My question is what do you do with the principal components? I have two principal components that explain 95% of the data. How do I use them for further data analysis or is the biplot the endgame with them?

Scree plot of iris data, pca

biplot

gung - Reinstate Monica
  • 132,789
  • 81
  • 357
  • 650
dataddodo
  • 21
  • 2
  • You might find this to be of interest: [Using principal component analysis (PCA) for feature selection](http://stats.stackexchange.com/q/27300/). – gung - Reinstate Monica Apr 23 '16 at 14:15
  • I corrected the spelling of "principle" to "principal". Shockingly, someone at Yale made the same spelling error http://www.yale.edu/ceo/Documentation/PCA_Outline.pdf . – Mark L. Stone Apr 23 '16 at 14:51
  • 5
    This is a strange question: why are you performing PCA if you have no idea what you want to do with it?? – whuber Apr 23 '16 at 16:32
  • 1
    Voting to close as too broad, following @whuber's comment. – amoeba Apr 28 '16 at 00:08
  • 5
    This seems like a reasonable question to me. The OP is clearly running PCA to learn more about PCA, statistics & R. A quickie overview of the uses of PCA seems like a reasonable thing to have on the site. – gung - Reinstate Monica Apr 28 '16 at 00:29

0 Answers0