From: jabba@_removeAntiSpam_dbs.dk   
      
   Hi.   
   I'm rather new to fuzzy clustering algorithms, and was starting to look at   
   clustering algorithms for classifying data sets into classes or clusters.   
      
   My problem is however that with fuzzy c and k means the number of clusters   
   needs to be known in advance, however i need to decide the number of   
   clusters.   
      
   If for example given 5 descriptions of houses:   
      
   h1(big, expensive, blue, 215 sq metres)   
   h2(big, expensive, blue, 213 sq metres)   
   h3(big, expensive, blue, 219 sq metres)   
      
   h4(small, cheap, green,  98 sq metres)   
   h5(big, cheap, green, 98 sq metres)   
      
   Intuitively h1, h2, and h3 could very well be the same house, as well as h4   
   and h5 could be the same house.   
      
   However the case is not always that there are c=2 or k=2 clusters, but the   
   amount of clusters must be set based on similarity among the clusters.   
      
   Does anyone have an idea of other forms of clustering algorithms or   
   technique, which could solve this problem. (I'm aware that i could make   
   several calculations with a range of values for the amount of clusters and   
   compare these, but maybe there is a better approach.)   
      
   Thanks in advance.   
      
    - Jeppe Moritz Hansen   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)