Online data mining services for dynamic spatial databases I: system architecture and client applications

Abstract

This paper describes online data mining services for dynamic spatial databases connected to environmental monitoring networks. These services can use Artificial Neural Networks as data mining techniques to find temporal relations in monitored parameters. The execution of the data mining algorithms is performed at the server side and a distributed processing scheme is used to overcome problems of scalability. To support the discovery of temporal relations, two other families of online services are made available: vectorial and raster visualization services and a sonification service. The use of this system is illustrated by the DM Plus client application and the SNIRH Data Mining Web site. The sonification service is described and illustrated in the part II paper.