iSixSigma

Trying to Identify the Distrubution

Six Sigma – iSixSigma Forums General Forums Tools & Templates Trying to Identify the Distrubution

Viewing 8 posts - 1 through 8 (of 8 total)
  • Author
    Posts
  • #55710

    Sambit Dash
    Participant

    Dear Friends Unable to identify the distribution for the bellow data. The data represents no of candidates submitting documents for a Background verification process within time periods. Each no represents cycle time of each candidate. EG if its 0 candidate takes 0 days to submit the documents.

    My task is to identify probability for 95% candidates submission within how may days.

    My assumption is its has to be a Poisson distribution as we are calculating no of success within a time period. 1-7 days 7-15 days etc.

    Please suggest.

    Data is attached in xls.

    0
    #201280

    MBBinWI
    Participant

    @sambit.ximb – sorry, no attached file.

    0
    #201281

    Sambit Dash
    Participant

    Re-attached….

    0
    #201286

    Sergey
    Participant

    I have no tools by hand to confirm whether it is Poisson or not (based on data seems like it is).

    What kind of risks do you have if you choose wrong distribution?

    0
    #201287

    Sambit Dash
    Participant

    I have to estimate Maximum likely hood for operation team to sign an SLA with client. Has to be correct.

    i analyzed manually in xls and in minitab. minitab suggest doesnt fit. xls shows poisson is closest possible distribution.

    0
    #201288

    Robert Butler
    Participant

    I did a quick graphical check of your data. Process Initiation and TMI approval are such that I don’t think you will go wrong treating them as Poisson. The issue are the BGV initiation and final. These two have a bi-modal appearance with 0’s forming one group and everything else forming the other group. You might want to consider treating these as zero inflated Poisson.

    0
    #201290

    Sambit Dash
    Participant

    Thanks a lot robert …..

    0
    #201292

    Chris Seider
    Participant

    I’m curious why the data isn’t the same in each column–# of observations.

    0
Viewing 8 posts - 1 through 8 (of 8 total)

You must be logged in to reply to this topic.