Abstract
The requirement of GIS community in information is in vogue. Indeed, GIS is often extremely useful in many fields such as sciences, health, business, and community services to do everything. This includes creating, querying, overlaying maps and analysis of the spatial and non-spatial information. To carry out all these processing, the GIS refers to the data collected and stored in the GDB. Seen the multitude of GIS users' preoccupations, there is a growing interest and awareness in reaching the valuable and strategic data by enriching the GDB. This is made possible through mining different data sources to complement the native GDB. In this context, this chapter deals with an approach devised to enrich the GDB semantic data by interacting reactive agents to process documents related to the geographic entities. It deals with a modular approach consisting of fragmentation, theme identification, delegation, filtering and eventually results refinement.
TopBackground
The GDB enrichment is becoming a necessity in front to the increasing demand on geographic information. The literature has revealed different efforts towards alimenting the GDB with complementary data to satisfy the majority of requests.
In this context, the GeoNode (Geographic News On Demand Environment) (Hyland et al., 1999, Clifton et al., 2001) operates on news-based information related to some geographic entities. This data source is to be investigated to pinpoint the embedded stories. The latter are explored by the MITRE’s Alembic system (Day, 1997) to pick out the named entities. A gazetter is used to geo-reference the places. The story datelines are also captured. The overall extracted data stands for the metadata. The named entities are explored to extract the topic via a clustering process (Clifton & Cooley, 1999). Hence, the co-occurrence of the names related to organizations, persons and places, many times, is considered as the description of a certain topic.