Parsing "grammatically incomplete" natural language queries to spatial databases

Of the applications of information system technology, geographic information management and processing can be categorized into the group that has both the most diversified users and the largest number of imprecise concepts. Formal query languages have limitations when applied to these fields. A natural language user interface may be more expressive in dealing with geographic information and can improve accessibility of geographic information systems (GISs). A challenge in creating a natural language user interface is that user queries in a natural language may be grammatically incomplete. That is, some grammatical constituents may be missing. The article reports on a technique for parsing such queries. This technique is based on the fuzzy set possibility theory, which is used to handle the uncertainty caused by missing constituents and ambiguity.