SPATIAL KNOWLEDGE EXTRACTION FROM GEOGRAPHICAL DATABASES An approach based on the Controlled English Query Language Geo-Q and Conceptual Graphs

Conceptualizing spatial relations through natural language seems to be closer to the model humans adopt for spatial reasoning. Controlled languages, which are subsets of natural ones, could be used in order to overcome the difficulties that emanate from the required experience in programming languages. In this paper, we propose a methodology in which questions are expressed in a new controlled query language called Geo-Q, for the purpose of extracting spatial knowledge from geographical databases. In the first step of the methodology, the question is parsed and a direct correspondence between the components of the question formulated in Geo-Q and the concepts/relations which constitute a Conceptual Query Graph is established. In the second step, the resulting Conceptual Query Graph is adequately transformed to an SQL query and the execution of the SQL query finally leads to the answer.