If you have a new account but are having problems posting or verifying your account, please email us on [email protected] for help. Thanks :)
Hello All, This is just a friendly reminder to read the Forum Charter where you wish to post before posting in it. :)

SPSS Errors - Hierarchical cluster analysis Ward's method

  • 24-08-2013 7:37pm
    Registered Users Posts: 514 ✭✭✭ bedrock#1

    Hi there,

    Trying to do a Hierarchical cluster analysis using Ward's method to extract media frames from a data set I've created in spss.

    I'm having a problem with the syntax file with spss showing me the following errors in the output file.
    >Error # 63 in column 14. Text: C:\DOCUME~1\Ross\LOCALS~1\Temp\spss4256\spssclus.tmp
    >The file does not exist.
    >Execution of this command stops.

    ERASE FILE='C:\DOCUME~1\Ross\LOCALS~1\Temp\spss4256\spssclus.tmp'.

    >Error # 2302 in column 13. Text: C:\DOCUME~1\Ross\LOCALS~1\Temp\spss4256\spssclus.tmp
    >The specified file does not exist and cannot be erased.
    >Execution of this command stops.

    I cannot find anything online that gives me a solution I can understand!

    If anyne has any idea how I can fix this I would be eternally grateful!!



  • Registered Users Posts: 3,483 ✭✭✭ Ostrom

    Are you running the procedure through the dialogs or through the syntax? Honestly, I have no idea what that message is - I use clustering quite a bit and I've not seen that (it's been a while since I've used SPSS). Are there missing values in the columns?

    I'm sure I sound like broken record saying this, but do you have access to stata, or would you entertain downloading r? The SPSS file is readily interchangeable with a little fiddling (especially easy with stata), and the clustering routines are much simpler in stata also. At a push, you could estimate the number of clusters using this rule of thumb: [sqrt(n/2)], then run a k-means analysis in SPSS, but I assume you want to estimate the grouping structure from the dendrogram?

  • Closed Accounts Posts: 2,980 ✭✭✭ Kevster

    If you're still stuck, then you can just use the following code in R Programming Language:

    plot(hclust(dist(MATRIX, method="euclidean"), method="ward"), labels=MATRIX$labels)

    ...where MATRIX is a matrix or data-frame of data values