To whom it may help me,

I know there are a number of Indices that may be used to “validate” 
a “hard cluster partition” or a fuzzy cluster partition”, and I also 
know that these indices depend on the number of points that each 
cluster may have. 

I am wondering up to what point is this dependence important!?.

In fact, and here is my question, can I use “cluster validity 
indices” when I only have 8 or 9 points (in a multi-dimensional 
space?. Does it make sense in this case to talk about “compacity”, 
or “density”?.

I would appreciate apart form your opinion,  some papers where I may 
see this discussion.

Also, all comments are welcome!.

Thanks,


Pedro Latorre.
           

----------------------------------------------
CLASS-L list.
Instructions: http://www.classification-society.org/csna/lists.html#class-l