- author = "Chris Gathercole and Peter Ross",
- title = "Dynamic Training Subset Selection for Supervised Learning in Genetic Programming",
- booktitle = "Parallel Problem Solving from Nature III",
- year = "1994",
- editor = "Yuval Davidor and Hans-Paul Schwefel and Reinhard M{\"a}nner",
- series = "LNCS",
- volume = "866",
- pages = "312--321",
- address = "Jerusalem",
- publisher_address = "Berlin, Germany",
- month = "9-14 " # oct,
- publisher = "Springer-Verlag",
- keywords = "genetic algorithms, genetic programming, DSS",
- ISBN = "3-540-58484-6",
-
URL = "
http://citeseer.ist.psu.edu/gathercole94dynamic.html",
-
URL = "
http://www.cs.ucl.ac.uk/staff/W.Langdon/ftp/papers/94-006.ps.gz",
-
DOI = "
doi:10.1007/3-540-58484-6_275",
- size = "10 pages",
-
abstract = "When using the Genetic Programming (GP) Algorithm on a
difficult problem with a large set of training cases, a
large population size is needed and a very large number
of function-tree evaluations must be carried out. This
paper describes how to reduce the number of such
evaluations by selecting a small subset of the training
data set on which to actually carry out the GP
algorithm.
Three subset selection methods described in the paper are:
Dynamic Subset Selection (DSS), using the current GP run to select difficult and/or disused cases,
Historical Subset Selection (HSS), using previous GP runs,
Random Subset Selection (RSS).
Various runs have shown that GP+DSS can produce better results in less than 20percent of the time taken by GP. GP+HSS can nearly match the results of GP, and, perhaps surprisingly, GP+RSS can occasionally approach the results of GP. GP+DSS also produced better, more general results than those reported in a paper for a variety of Neural Networks when used on a substantial problem, known as the Thyroid problem.",
-
notes = "PPSN3 Describes how to reduce the number of fitness
case evaluations in difficult GP problems by selecting
a small subset of the training data. Dynamic Subset
Selection can produce better results than GP in less
than 20percent of the time. Population size of 5,000
and 10,000.
DSS used in Discipulus \cite{francone:manual}",
