Abstract
This paper investigates the use of genetic programming
(GP) for learning of pattern recognition programs. The
central topic here is the introduction of GP
incorporating partial order of solutions as opposed to
the standard complete (linear) order imposed by the
scalar fitness function. We claim that such an
extension protects the `interesting', however worse
w.r.t. the value of the fitness function, solutions
from being discarded in the selection process, and thus
increases the diversity of the population. That
hypothesis is verified on a real-world case study
concerning the recognition of handwritten characters.
Users
Please
log in to take part in the discussion (add own reviews or comments).