K-MEANS CLUSTER ANALYSIS

From: Stefania Kalogeraki (S.Kalogeraki_at_reading.ac.uk)
Date: 10/14/04 

Date: Thu, 14 Oct 2004 12:38:45 +0000 (UTC)

Hi,
   I conduct a comparrative research between 3 countries for drug
taking among students. I am trying to create groups of drug users
based on their frequency of use. My initial data were ordinal (e.g
1=never tried, 2=1-2 times, 3=3-5 times etc) and I transformed them to
interval : never tried=0, 1-2 times= 1.5, 3-5 times= 4 etc)in ables to
conduct k-means cluster analysis. My data set is rather big almost
8000 students. I haven't properly understood how exactly do I
exactly achieve the maximum efficiency described in the "help" of
spss. OK ,I take a sample of cases and use the Iterate and classify
method to determine cluster centers. Is the number of clusters
subjective? Do I choose the final clusters of these results? And do I
click the "option of membership" so automaticly spss will creat a
column with the relevant clusters? Cause I really can't understand how
this column of clusters will appear to the sample of the data set. So
the file which will be included in "Write final as File" will be the
result of the clusters from the sample?
It's a bit of a mess in my mind!
Thanks in advance for your help
S

Relevant Pages

  • Cluster Analysis Suggestions
    ... 1,400 college students. ... My problems is the number of clusters to retain. ... model each of the cluster solutions by the original set ... scree plot on goodness of fit statistics (log-likelihood, chi-square, or ...
    (sci.stat.edu)
  • Re: K-MEANS CLUSTER ANALYSIS
    ... > taking among students. ... I am trying to create groups of drug users ... Do I choose the final clusters of these results? ...
    (sci.stat.math)
  • Re: spatial autocorelation methods
    ... > a 'detect' in the other data set). ... If the real potential sum is ... Developing clusters is a different thing than correlation I should think. ... As far as lateral movement I meant if the satellite images are of the same ...
    (sci.image.processing)
  • Re: Computational complexity of cluster detection
    ... computational complexity of computing whether a data set contains one ... of more clusters is well below O, possibly O, if one uses ... many problems for which solving ... distributions for the complete data set. ...
    (comp.theory)
  • R clustering using diana and Calinsky and Harabasz Index
    ... managed to create dendrogram for this data set using diana() in R, ... however this only gives me the tree and not the clusters themselves. ...
    (sci.stat.edu)