文档介绍：A ic Programming Framework for Two Data Mining Tasks:
Classification and Generalized Rule Induction.
Alex A. Freitas
University of Essex
Dept. puter Science
Colchester, CO4 3SQ, UK
******@
ABSTRACT submitting SQL queries to a database server, so achieving a
tight integration between GP and relational databases. The
This paper proposes a ic ability of GP to search many parts of the program (database
query) space in parallel and the robustness of GP render the
programming (GP) framework for two DM system more autonomous, minimizing the need for
major data mining tasks, namely domain knowledge to guide the search. Finally, efficiency
classification and generalized rule can be significantly improved by using a parallel SQL server
induction. The framework emphasizes to evaluate the fitness of an individual. Note that the use of
the integration between a GP algorithm this kind of database server does not require any modification
in the GP framework - . the SQL queries generated by the
and relational database systems. In GP are automatically parallelized by the SQL server.
particular, the fitness of individuals is There are several kinds of DM tasks [Fayyad et al. 96],
computed by submitting SQL queries to depending mainly on the application domain and the user
a (parallel) database server. Some interest. Some of the major DM tasks