Hi, I'm a little new to data mining and would definitely appreciate some tips.

I'm using clustering algorithms looking for possible grouping in some variables described below.

I've been using the excel data mining add-in which connects to SSAS and uses the EM algorithm by default. I'm also using R, so far with the Kmeans algorithm.

I have two independent variables X, and G which are integers

and three dependent variables A, B, C which are related to X and G by the equations below:

A=(((2xG)x(0.06)+K)/((2xG)x(.4)+(2xG)x(0.06)+K+X)))x100

B=(X/((2xG)x(.4)+(2xG)x(0.06)+K+X))X100

C=100-((((2xG)x(0.06)+K)/((2xG)x(.4)+(2xG)x(0.06)+K+X)))x100 + (X/((2xG)x(.4)+(2xG)x(0.06)+K+X))X100)

where K is an arbitrary contant which varies.

Most of the data follows the formulas above with some variation.


I'm using clustering algorithms to look for groupings in the A,B, C variables.

Any advice on strategies for looking for clusters, and how to tell when I've been successful would be greatly appreciated.