iSixSigma

How to normalise distribution free data

Six Sigma – iSixSigma Forums Old Forums General How to normalise distribution free data

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
    Posts
  • #47638

    Ron Brown
    Member

    I have been working with data that are ditribution free and contains zeros. I was trying to normalize them before conducting a capability analysis, but I could not use the box Cox transformation nor could I use the Johnson transformation.
    Doe anyone has an idea?
     

    0
    #159027

    annon
    Participant

    Use a capability metric that doesnt require normality:  DPU or  DPMO
    Use Individual Distribution Identification in MTB to determine your underlying distribution type and then run the MTB non-normal capability assessment accordingly.

    0
    #159028

    Mikel
    Member

    Why do you think your data is distribution free?

    0
    #159029

    Ron Brown
    Member

    I used Minitab to do an “Individual Distribution Identification test” and I used the option “Use all distributions” and the data did not fit in any one of them.

    0
    #159030

    Mikel
    Member

    If you are truly distribution free (very unlikely), there is no transform that you can do.
    It is more likely that you have data from a poorly controlled process. Go do some simple standardization before you worry about any analysis of data.
    What kid of a process is the data from?

    0
    #159037

    Joe BB
    Participant

    What kid of a process is the data from?
    You will never change.

    0
Viewing 6 posts - 1 through 6 (of 6 total)

The forum ‘General’ is closed to new topics and replies.