Intelligent agents in managing datamining processes: predicting probability of default of bank customers

Povzetek

Business Intelligence (BI) systems are heterogeneous by their nature, while all plans to unify them eventually fail. Moreover, R&D efforts in building proactive BI let us conclude that the heterogeneity may become chaotic. The especially vibrant is the Data Mining R&D field, where new algorithms and methods emerge. We claim that, for mastering such a heterogeneous and dynamic system, a programmatically accessible and self-managed knowledge management system, capable of storing and representing the whole content of a BI system is required. The properties of Knowledge Management Structures (KMS), combined with Intelligent Agents (IA) acting in Multi-Agent Systems (MAS) show great capabilities, and have, in limited scope, proved their strengths, acting in heterogeneous systems. There are plenty of research papers, handling new technologies of IA and KMS, largely focusing on a single point of research in its research field and we found a great lack in research reports, trying to combine some of technologies or even put them to a test or a real-life scenario. In this paper we present the results of a research project, where the potentials of MAS and KMS in managing complex BI environments were verified. The research project involved building a prototype of a system BIMAS, capable of managing data mining processes, using MAS and KMS technologies. The prototype was tested for its behavior in a real case of predicting the probability of default (PD) of commercial bank customers. The results of the prototype BIMAS are discussed in two stages: The current technology state of research and the open research issues are discussed, providing some promising results, enabling the technology to better adapt to the changing environment. On the user side the results are evaluated, giving new requests for the technology to cope with. The research provides the positive answer to the research questions on usability of the MAS and KMS properties in a BI environment, proving it's potentials as very useful. It also provides the better understanding of BI needs and displays properties of new technologies which need further developments.

Ključne besede

poslovna informatika;management;vodenje;znanje;bančništvo;

Podatki

Jezik:	Angleški jezik
Leto izida:	2008
Tipologija:	1.01 - Izvirni znanstveni članek
Organizacija:	UM EPF - Ekonomsko-poslovna fakulteta
UDK:	659.2:004
COBISS:	9764892
ISSN:	1313-2555
Št. ogledov:	1131
Št. prenosov:	30
Ocena:	0 (0 glasov)
Metapodatki:

Ostali podatki

Sekundarni jezik:	Angleški jezik
Sekundarne ključne besede:	poslovna informatika;management;vodenje;znanje;bančništvo;
URN:	URN:SI:UM:
Strani:	str. 544-562
Letnik:	ǂVol. ǂ2
Zvezek:	ǂiss. ǂEconomy & Business
Čas izdaje:	2008
ID:	8717467