Title :
SPOTLIGHT: a data explanation system
Author :
Anand, Tejasvi ; Kahn, Gregory
Author_Institution :
A.C. Nielsen, Bannockburn, IL
Abstract :
SPOTLIGHT is a knowledge-based product that makes it possible to extract information from large databases of point-of-scale scanner data. SPOTLIGHT can be classified as a data explanation system. Compared to simpler data interpretation systems, in data explanation systems large bodies of data are analytically relevant; explanation models are complex, regarding the causal interrelation of events over time; and user needs in terms of both analysis and presentation are heterogeneous. The authors illuminate some of the design challenges these systems face by discussing the requirements, functionality, and system architecture of SPOTLIGHT. These challenges include the need to accommodate a broad spectrum of tasks and user roles; to represent generic knowledge while allowing user-specific customization; to integrate with third party software; and to achieve effective use of centralized mainframe and distributed PC computing resources
Keywords :
explanation; knowledge acquisition; knowledge based systems; mark scanning equipment; point of sale systems; POS; SPOTLIGHT; causal interrelation; centralized mainframe; data explanation system; design challenges; distributed PC computing resources; explanation models; generic knowledge; knowledge-based product; large databases; point-of-scale scanner data; system architecture; third party software; user needs; user roles; user-specific customization; Computer architecture; Data analysis; Data mining; Databases; Distributed computing; Information analysis; Knowledge based systems; Marketing and sales; Performance analysis; Software tools;
Conference_Titel :
Artificial Intelligence for Applications, 1992., Proceedings of the Eighth Conference on
Conference_Location :
Monterey, CA
Print_ISBN :
0-8186-2690-9
DOI :
10.1109/CAIA.1992.200003
