Abstract
gradient descent search in genetic programming (GP)
for object classification problems. In this approach,
pixel statistics are used to form the feature terminals
and a random generator produces numeric terminals. The
four arithmetic operators and a conditional operator
form the function set and the classification accuracy
is used as the fitness function. In particular,
gradient descent search is introduced to the GP
mechanism and is embedded into the genetic beam search,
which allows the evolutionary learning process to
globally follow the beam search and locally follow the
gradient descent search. This method is compared with
the basic GP method on four image data sets with object
classification problems of increasing difficulty. The
results show that the new method outperformed the basic
GP method on all cases in both classification accuracy
and training time, suggesting that the GP method with
the gradient descent search is more effective and more
efficient than without on object classification
problems.
Users
Please
log in to take part in the discussion (add own reviews or comments).