Building Standardization Model Of Address Based On Statistical Methods

Along with the speeding up of the informatization process, all enterprises and institutions, government functional departments urgently need to integrate the non-spatial data and spatial data to realize non-spatial data visualization which is used for daily work decision-making. In order to integrate non-spatial information and spatial information, we need to establish the connection between them, and geocoding is the important means to realize this connection.The purpose of geocoding is endows address location geographic coordinates (e.g., a street address). In real world, because the complexity and non-normalization of address data, so before conduct of geocoding work, we need to build address standardization model, we only then may match address data and thus endowing geographic coordinates. At present, the main problem which perplex Chinese geography encoding progress is the lack of standardized address model.This thesis came up with the concept of unit, address level and the level address. Using the statistical method establishes the statistic X and statistic Y. Using statistic X judge and cut out address keywords. Using the discriminance and statistics Y to identify the secondary address authenticity, and then encode all levels of address, on the basis of these establish a standardized address model.