Re: classification

[http://dummy.us.eu.org/robert (Robert)]

 > From: Leonid <http://www.gmail.com/~l>
 > Date: Tue, 3 May 2005 14:28:50 -0400
 >
 > Hi, Robert -- need an advice. I talked to a company about a prosect of
 > wroking on their image processing functionality. Seems like I made an
 > expression of more a "classification" rather than an "imaging" guy.
 > Now here's the rub, they quite unexpectedly sent me a long file with
 > numbers asking to classify it. I'm going to look at it and "try
 > things". What would be first things you'd try? Any handy SW?
 > 
 > Leonid

Handy software.  No.  I think Oracle has some hooks to do some simple
pattern matching.  And there's probably something out there on the 'net to
do this.  Perhaps even that package, Torch II, has something which works
on rows of data.

Since this is for "research purposes" only, you might OC1.  Search for
"OC1" on google.  OC1 was what imspired Marc's program.  It's an induced
decision tree.  There may be a way to format the input to its
requirements.

Other than that, as I said above, there are dozens of neural net and other
types of programs which does stuff.  A decision tree is easier to
understand, but, if you'd like to baffle them with bullshit, a neural net
may fit the bill.