Data Analysis not Completing

Home Forums PATN and Pattern Analysis PATN Discussion Data Analysis not Completing

  • This topic is empty.
Viewing 3 posts - 1 through 3 (of 3 total)
  • Author
    Posts
  • #421
    BEUKERM
    Member

    I was an occasional DOS PATN user who is making his first jump into PATN 3.03. It looks great. Love the online help. I have great hopes.

    Right now though I cannot get past the Data Analysis.
    When I use the default Gower Metric it will commence, gets to about the 20th iteration and drops out without an error message. When I use flexible UGPMA (default beta -0.1) I get “- ROW CLASSIFICATION FAILED”.

    I have spent the day checking and re-checking my data. This will probably mean that someone will find a simple error in it and I will feel very foolish.

    My data is vegetation classification data, litoral rainforest sites from Gippsland and Far South Coast NSW, 486 taxa (variables), 251 sites (objects). Imported from both MSExcel and .csv formats and transposed before attempting Data Analysis.

    I have tried reducing label lengths, putting quotation marks around labels and removing spaces in labels. As well as runing data analysis on un-transposed data. The problem remains.

    Also noticed the advise box on the bottom right does not swap the number of object and attribute around after data is transposed. I presume it should.

    Anyone with advise or help will be greatly appreciated. I apologise in advance for any oversight on my part. I am but a try-er.

    #494
    BEUKERM
    Member

    Hang-five on this request. Please ignore my posting for now. I realise now that I am in a two step process and will have to export and re-import my grouped data before I can go on with data ordination

    Love the history option.
    Hope I have not wasted anybodies time.

    😳

    #495
    lee
    Keymaster

    Hi Max

    Sounds odd, but from your last posting, I’m not sure where you are at.

    The Gower metric may fail for a number of reasons, most common is a zero ranged variable. If this happens, it should tell you, and as there won’t be a dissimilarity matrix, no classification or ordination is possible.

    Your right about the rows/cols in the status bar. This has been corrected in 3.04 – we are loading this to the site this week.

    Lee

Viewing 3 posts - 1 through 3 (of 3 total)
  • You must be logged in to reply to this topic.