Karan, Without knowing why clustering didn't work, it's hard to say what a better approach would be. Any other information you can give about the problem you're working on would probably help, too. In particular, how did you come up with your four categories? Typically, categories are not defined directly in terms of the data, because the problem is to find the relationship between the categories and data.
Matt On Wed, May 8, 2013 at 5:44 AM, Karan <[email protected]> wrote: > Hi All, > > I have some numerical data in pairs say X & Y and I want to divide(cluster, > may be) into four groups as LowX-LowY,LowX-HighY,HighX-LowY & HighX-HighY. > I > tried with clustering but unable to identify clusters(and i think is not > the > best way to achieve it). Can someone suggest any good(non-trivial) approach > to proceed. > > Thanks, > Karan > >
