Re: another attempt

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

 > From: Leonid <http://www.gmail.com/~l>
 > Date: Tue, 3 May 2005 17:00:34 -0400
 >
 > I got confused because the top of the message looked strange (I paste
 > it at the bottom).
 > 
 > I'm confused by the assignment. I have some bits and pieces of the SW
 > I wrote, but usually classification problems have some "context". In
 > some sense this may still be a legitimate problem.

It does make sense for, for example, a decision tree.  It becomes pretty
easy to classify cases (rows) used for the training.  With neural nets,
linear or logistic regression, or GAs, it is usually not obvious because
you have to stick values into a calculator to see if the classification is
working correctly or not.  Perhaps they just want a huge formula which
will predict a particular column given the other two columns.  If you
show 'em 3 huge formulas or neural networks, for predicting each of the
columns given the other two columns, perhaps that would "wow" them?

 > Did you ever get
 > anything like this as a test for employment???

No.

 > L