Neighborhood modeling algorithm of geographic entity in natural language geocoding

Neighborhood Modeling lays the basis for natural language geocoding with neighborhood properties. With the deep analysis of research results of former scientists, and the special requirement of Geocoding, Observation scale, geometry shape of geographic entity and spatial distribution are selected as the key influencing factors for determining neighborhood range of geographic entities in natural language address. Then, based on consideration of above four key influencing factors, we model the neighborhood of geographic entity based on fuzzy set theory, and design several investigation tables about human beings neighborhood range cognition of singe line or face shape entity and multiple entities etc. According to the analysis of the test result, several formulas and algorithms were put forward to calculate the buffering distance of neighborhood range of different type of geographic entity. It is concluded that the neighborhood modeling algorithm proposed in this paper can be applied in future geocoding algorithm, because it faithfully fits cognition of human beings about the neighborhood concept in natural language address.