Research On Fuzzy Chinese Address Matching Method For National Economic Census Application

Geocoding is a basic GIS technology, which is a process of giving the spatial location information to the natural language described address and locating in the map by a series of processing operations, including address standardization, word segmentation, database matching and spatial location. With the development of GIS, there are more and more industries demanding geocoding, such as public health, crime analysis, political science, disaster management, traffic forecasts and other fields. Foreign geocoding technology has matured and gradually moves towards a market-oriented industrialization. However, due to the differences in national circumstances, existing technologies from abroad can not be applied to our country directly. Therefore, further research needs to be done on the Chinese geocoding technology.The paper designed the experiment by using the economic census data of Beijng, and developed the geocoding tools at last. In the course of the study, the paper mainly set focus on four aspects, including: (1) Because of the low accuracy and low matching rate in geocoding of fuzzy addresses, this paper presents a rule-based Chinese address geocoding method, which improved the fuzzy address matching success rate by adding rule tree and ambiguity storage mechanism to the algorithm.(2) In the traditional process of geocoding, address segmentation and database matching are two independent steps, which resulting in the excessive access of the database and low system efficiency. Therefore, in the rule-based Chinese address geocoding method proposed in this paper, it combined the two processes into one, and accomplished the process of database matching at the same time of finishing the address segmentation, which improved the speed of geocoding.(3) The paper made some research and revision about the existing address model. According to the difference between administrative divisions part and detailed street address part, the paper separate the address into two parts,which improved the matching speed. (4) In order to reduce the workload of data collection and address standardization, the paper made effective use of economic census data by data mining, established the standard address dataset and completed the geocoding task in the project.