+ Reply to Thread
Results 1 to 4 of 4

Thread: PCA - To check

  1. #1
    Points: 230, Level: 4
    Level completed: 60%, Points required for next Level: 20

    Posts
    1
    Thanks
    0
    Thanked 0 Times in 0 Posts

    PCA - To check




    Hi
    I got a data set and tried my hand in doing PCA using SPSS and for the same set of data carried out PCA by using matlab program (first standardising data and then using princomp). I have got the PC1 Vs PC2 Vs PC3 plot to be different in both the cases. Can some one tell me which is correct. I can give the data for you to try. Please help.

    Thankyou all in anticipation
    Amutha Daniel

  2. #2
    Ninja say what!?!
    Points: 8,297, Level: 61
    Level completed: 49%, Points required for next Level: 153
    Link's Avatar
    Posts
    1,165
    Thanks
    37
    Thanked 84 Times in 76 Posts

    Re: PCA - To check

    I don't understand what you're trying to say. PCA is simply a method that will try to explain as much of the variance in your data as possible with a new set of variables, normally in decending order of variance captured. Thus, the PC1 will explain the most variance in your data, followed by PC2, and so forth.

  3. #3
    Points: 222, Level: 4
    Level completed: 44%, Points required for next Level: 28

    Posts
    5
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: PCA - To check

    I'm not an expert but there is a difference between princomp and prcomp in 'R', because one uses eigenvectors and the other uses SVD. Maybe the difference you see here relates to something like that, and gives slightly different PC scores

  4. #4
    Beep
    Points: 60,883, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    Discussion EnderPosting AwardCommunity AwardMaster TaggerFrequent Poster
    Dason's Avatar
    Location
    Ames, IA
    Posts
    11,034
    Thanks
    259
    Thanked 2,132 Times in 1,814 Posts

    Re: PCA - To check


    I don't know the specifics each algorithm is using but there is the added component that you can either use the covariance matrix or the correlation matrix when doing finding the PCs and it makes a difference if your data isn't already standardized properly.
    Morte a tutti i raptors
    001100010010011110100001101101110011

+ Reply to Thread

           




Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats