import matlab .mat

devl82devl82 Member Posts: 4 Contributor I
edited November 2018 in Help
Hello ;D i just started using rapidM.. very happy till now:)
just one question:
is it possible to import .mat files from matlab?
<i have a 32x786432 matrix which when i save as .txt (>40MB) cant be opened from rapid:<>
thnks ;D

Answers

  • landland RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 2,531 Unicorn
    Hi,
    if you have exported it in a proper .csv style, there should be no problem. Perhabs you have transposed the matrix in the wrong way? If I read the numbers correctly, you have 32 examples with 786.432 attributes? That would be probably the worst proportion for learning I've ever heard including my experience in the life sciences...

    If that's not the case, did you try the data configuration wizard of the ExampleSource operator?

    Greetings,
      Sebastian
  • devl82devl82 Member Posts: 4 Contributor I
    Hello again ;D
    Thank you for your quick response and soorry for my late reply:):)
    My dataset has 32 attributes not 32 instances, so after i transposed the matrix i imported it to rapidminer successfully:) thnks again:):)

    ps i am trying to find out how attributesWeights are calculated with the Feature selection operator.. Cant find an explanation iin the documentation propably because its a very standard/simple procedure. A little hint plz??
  • landland RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 2,531 Unicorn
    Hi,
    the feature selection operator removes some of the attributes from the example set and keeps the others. All attributes kept recieve weight 1 and all others weight 0, which will cause them to vanish on applying the weights later on.

    Greetings,
      Sebastian
  • devl82devl82 Member Posts: 4 Contributor I
    Hello again:) Its so nice that you find time to answer these simple questions:):)
    I do know that feature selection can turn on and off features based on some criteria depeding on the method. I asked about weights because when running your example of WrapperXvalidation (with the feature selection procedure -also in the tutorial), the output is a vector containing weights for each feature not zeros and ones. These weights how are calculated?? (if its a long exmplanation just send a link)
    thnk you
  • landland RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 2,531 Unicorn
    Hi,
    as far as I know the attribute weights are avaraged over all runs of the WrapperValidation...

    Greetings,
      Sebastian
Sign In or Register to comment.