1

I'm using PCA for the first time and just experimenting with it. I used PCA on my dataset that can be found here 

wine <- read.csv("wine_nocolor.csv")
wine1 <- wine[2:13]
pca <- princomp(wine1[1:11], cor=TRUE)
biplot(pca)

produces this plotenter image description here

Questions

  • How would I interpret the above graph?
  • My aim is to apply dimension reduction to see which attributes can be taken out while not losing meaning of the data. How can I do this with PCA?
birdy
  • 481
  • 8
  • 14
  • My guess is that this plot is the projected data in 2D where the green arrows are also *projected* directions of maximum variance found by PCA. – Vladislavs Dovgalecs Apr 03 '15 at 06:55
  • This question should be closed: there is many similar questions already on the site. Search `PCA biplot` and read half a dozen top relevant items. – ttnphns Apr 03 '15 at 08:31
  • 1
    -1. This question shows zero research effort. – amoeba Apr 03 '15 at 19:36
  • @amoeba does the duplicate also show no effort? I've tried to generate the plots for my dataset, I just need some help in making sense of them. – birdy Apr 03 '15 at 19:43
  • The duplicate sounds more reasonable :) I have just left a comment there. There was indeed some overlap between these two questions, it is usually better to avoid that. E.g. ask one specific question, wait until it (hopefully) gets clarified, then with your increased understanding write a follow-up question, etc. – amoeba Apr 03 '15 at 19:45
  • Thanks. I just didn't want it to seem like I'm not doing any effort from my end. I'm finding all this information very exciting but just lack the statistical background required to make sense of it all. – birdy Apr 03 '15 at 19:46

0 Answers0