Due to recent updates, all users are required to create an Altair One account to login to the RapidMiner community. Click the Register button to create your account using the same email that you have previously used to login to the RapidMiner community. This will ensure that any previously created content will be synced to your Altair One account. Once you login, you will be asked to provide a username that identifies you to other Community users. Email us at Community with questions.

"Retrieving original data after PCA"

atharkharalatharkharal Member Posts: 3 Contributor I
edited June 2019 in Help
I use PCA to retain 95% variance. I also get a pre-processing model for ApplyModel but it does not give me original variables responsible for 95% variance. Instead I get a listing of PC-1 etc etc. Help please.
regards
Tagged:

Answers

  • MariusHelfMariusHelf RapidMiner Certified Expert, Member Posts: 1,869 Unicorn
    Hi,

    if I get you right you want to see something like an attribute ranking? Maybe then you want to have a look at the Weight by PCA operator. If that is not what you are looking for, please explain a bit more detailed what you are trying to do.

    Best regards,

    Marius
  • atharkharalatharkharal Member Posts: 3 Contributor I
    Thank you for the response. My problem still not solved. Let me give an example:
        Suppose I have 4 regular attributes viz. A1,A2,A3,A4 and 1 label viz L.  Now after running PCA, it gives me eigen values and vectors which tell that only two principal components PC_1 and PC_2 contribute 99%. For me this is not enough for me. I want to see those two attributes, say, A1 and A3 which are 99% responsible for the variation. Help, required.

    regards,
  • earmijoearmijo Member Posts: 271 Unicorn
    I think there is some confusion here. Principal components are linear combinations of the original attributes. Unless the loadings are unit vectors there is no equivalence between a principal component and the original attributes. Rapidminer gives you the components (which are combinations of the original atts). Am I understanding incorrectly your question?
Sign In or Register to comment.