Natural Language-based Spatial Data Retrieval

Spatial database has achieved a great of development which has been widely used in geographic information science, mechanics CAD, image database, environmental protection, city planning, and so on. Especially along with the development of the technology of the storage and the index of spatial data, the performance of the query of spatial database has been greatly improved.On the contrary, the requirements of information processing aren't completely fulfilled yet. The desire of using natural language to query the spatial database is stronger and stronger, and prompts the combination of the spatial database technology and the natural language understanding technology.In order to solve the problems mentioned above, the major contributions of the paper are summarized as follows.First, the model is presented that uses xml to describe the spatial data stored in the object-oriented database. The paper had also analyzed the indexing methods of spatial database, the R-Tree index style, the R~+-Tree index style and R-Tree index style are studied in detail.Then, the paper researches several key technologies about natural language understanding, including morpheme analysis, syntax analysis and semantic analysis. The WordNet dictionary is used in the morpheme analysis. Syntax analysis analyzes the relationship between words. At last the paper uses the semantic analysis to understand the natural language input by user.Finally, the system framework is designed based on the technologies above. There were two major parts in the framework. One is the natural language understanding, in which the synset is made by using WordNet to scan the data-table names and field names of the database. Then, the paper works out the rules and template to convert the natural language to standard SQL to query the database. And the other part is the design of special database, including the design of create, insert, update, delete algorisms of R-Tree-Index. And the paper designed several spatial...