Research on a Chinese word segmentation method based on spatial information query of GIS

Abstract

Spatial information query is one of the most fundamental applications in GIS. Spatial information query based on Chinese means making a query in forms of Chinese language. Word segmentation is the first step in the process of query understanding. With the help of related searching method to extract the characteristic words from the input sentence, we can segment it into several parts for further analyzing, and then map them to a subset of spatial information database and feed back the result to the user. This thesis tries to improve the study of Chinese spatial information query and proposes a word segmentation method based on a set of characteristic words. The method rearranges the query sentence according to the given patterns and abandons the irrelevant words. Moreover, when meeting the wrong object names or unknown ones, it will amend or insert them into the dynamic library automatically. Such word segmentation method not only takes away the identification of irrelevant words but improves its efficiency of spatial information query and makes for the following mapping to the spatial database.