One issue is whether you can even rationalize a non-linear or
discontinuous answer, should one fit your data perfectly and predict new
data accurately...
What symbolic regression adds over ordinary [linear] regression is the
ability to *find* programs (equations if you like) that involve software
constructs that aren't analytically tractable. Loops, recursive
functions, complicated conditionals, etc.
(I haven't used Eureqa per se, but I have implemented genetic
programming algorithms.)
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at http://www.friam.org
