Towards Data-Driven Hybrid Composition of Data Mining Multi-agent Systems

Autoři: Mgr. Roman Neruda, Ph.D.
Rok vydání: 2009
Typ publikace: Konferenční příspěvek (mezinárodní konf.)
Periodikum: Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing
Projekty: Centrum aplikované kybernetiky
Bibliografická citace: NERUDA, Roman. Towards Data-Driven Hybrid Composition of Data Mining Multi-agent Systems. In: LEE, R., ISHII, N., eds. Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing. Berlin: Springer, 2009, S. 271-281. Studies in Computational Intelligence, 209. ISBN 978-3-642-01202-0
Dostupné z: doi:10.1007/978-3-642-01203-7_24
DOI: 10.1007/978-3-642-01203-7_24
ID: 328423

Abstrakt

The approach of solving data mining tasks by a collection of cooperating agents can profit from modularity, interchangeable components, distributed execution, and autonomous operation. The problem of automatic configuration of agent collections is studied in this paper. A solution combining logical resolution system and evolutionary algorithm is proposed and demonstrated on a simple example.