Rule-based Chinese Address Segmentation And Matching Methods

In the information age of today, there are a large number of address information in the city departments. Most of the datas are non-spatial information, we can not share them by Geographic information system. So it's a main part of digital city building to informationize the city address information.Geocoding is a method to informationize the city address information, which provides a way of translating the text address to geographic coordinates. By geocoding technology, a large number of socio-economic data will become spatial information in the form of coordinates, data sharing can be achieved between the city departments and the industry, so there will be a more rapid and effective spatial analysis and decision-making.The paper takes the research of the WuHan addresses as the project background, using the address Geocoding technology to achieve address rapid query and socio-economic data spatialization, and build address database. Then we can share information in the city departments.The main contents of the research are:(1) Improving the existing address model, then we made address standardization by the new model, and build the standard address database.(2) Research some address segmentation and geocoding methods, and propose a rule-based Chinese address geoeoding method. We add rule tree and ambiguity storage mechanism to improve the success rate of fuzzy address matching.(3) The paper creates a learning system, so we can add database with fail-matching address by address adding module.