iSixSigma

Normal and non-normal data

Six Sigma – iSixSigma Forums Old Forums General Normal and non-normal data

Viewing 5 posts - 1 through 5 (of 5 total)
  • Author
    Posts
  • #48704

    Jose Ramon
    Participant

    Hello.  Six Sigma Professionals
    I have one inquiry:
    What happen in this case you have non normal data …when I transfom this data using Box-Cox and this cannot make it out. When I calculate Do I have to make the capability analysis for non-normal data ?.If this is the case there are lot of choices : weibull, exponencial, etc.  How can I decide the best one ?
    One more question:
    1. If your data you transformation with lamda X is good, this mean you transform your data to perform you capability analysis.  What happen with your LSL and USL that you have ?
    For instance.  LCL = 0  UCL = 300
               Time (sec)                               Transformation with lamda 0.5
               39                                            6.24
               65                                            8.06
               12                                           3.46
    of course the data is not just 3…it is 100 but this is the concept .
    Please let me know. Thanks so much
    Jose Ramon

    0
    #164992

    Robert Butler
    Participant
    #164996

    Bart Simpson
    Participant

    Hi Jose
    You should try  “Graph > Probability Plot” in Minitab in order to see your data set fits well to which probability distribution.
    B.S.
     

    0
    #165005

    Ward
    Participant

    If you have Minitab 15,
    Stat>Quality Tools>Individual Distribution Identification also works.

    0
    #165006

    Ward
    Participant

    What is the purpose of transforming the data? If it is to determine process capability and you have at least five good and at least five bad, then just determine the percentage good and you have the yield.
    If you must transform the data because you have less then five good and five bad, then whatever lambda value you use for the original data, must also be used for the spec limit(s). In your example, you chose lambda = .5 (square root) so you would also have to take the square root of the spec limit(s).

    0
Viewing 5 posts - 1 through 5 (of 5 total)

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