Karan, The clusters that come out of the KMeans algorithm won't necessarily fit the categories that you've defined. It all depends on the data. For example, the clusters could all be in a line (like at (6,6), (3,3), (-3, -3), and (-6, -6)), so there's no cluster that fits the description "High X-Low Y". I'd recommend just plotting the cluster means (and data, if there's not too much of it) and seeing what it looks like.
Matt On Tue, May 7, 2013 at 1:09 PM, Karan <[email protected]> wrote: > Hi All, > > I am doing clustering of some numerical data using KMeans algorithms(k=4). > My data consists of 2 decimal values[X,Y], let's say X&Y and expecting them > to be clustered into 4 clusters as LowX-LowY,LowX-HighY,HighX-LowY & > HighX-HighY. Though my 4 clusters are formed but I am unable to identify > which cluster belongs to which category. I can do it by mathematical > computations outside the logic but would be good if algo identifies it for > me. > > Thanks, > Karan > >
