Unordered attributes

obben

I have inputs with individual properties. For example XXXX(a,b),YYYY(a,c). I don't have  a clear label from inputs, but it is another set similar to inputs (such as XXXX(b,c),ZZZZZ(e,f).

For any given input  set, I would like to predict what the output set should be. I think some kind of  probabilistic learning would work.

There could be multiple input elements with same name but  different properties (for example XXXX(a,b) and XXXX(a,k). The input is unordered so I cannot have a big input vector.

Is it possible to handle this with rapid miner?

Thanks in advance.