Re: data mining of subsets

I looked up latent semantic indexing (LSI) and it does not seem very similar. The strings in my case are identifiers (names) and have no internal structure (well almost none). Each string has a lot of context associated with it, like time/date it was requested, the properties of the request and who made the request. I'm being vague here about what is actually going on but the specifics should not be necessary as far as data mining is concerned. A similar approach to LSI might be useful though. The rows could be all possible strings and the columns all possible values of the string parameters. But I would much rather (and will probably only have time for) use some ready made tool that I can apply to my problem.

Max Received on Fri Jul 27 2007 - 03:49:20 CEST