重庆市中文语义地址模型构建方法探讨  

Construction Method of Chinese Semantic Address Model in Chongqing

作　　者：臧英斐 王斌[2] 瞿晓雯 

机构地区：[1]重庆交通大学土木建筑学院,重庆400074 [2]重庆市地理信息中心,重庆401121

出　　处：《地理空间信息》2015年第3期122-125,12,共4页Geospatial Information

基　　金：测绘遥感信息工程国家重点实验室开放基金资助项目(13R03);重庆市教委科技资助项目(KJ1400325);重庆交通大学博士基金资助项目(2012kjc2-011)

摘　　要：现有重庆市地址模型存在地址不完整、歧义、口语化等弊端。以南岸为例,针对现有数据特点,围绕地址数据库建设及地理编码的需求,归纳非结构化中文地址的特点,优化现有地址要素结构。提出了基于中文语义的地址模型构建方法,建立要素间拓扑关系,增加上下文约束力,引入支持向量机,有效避免中文自然语言表达歧义,提高地址解析的准确度和效率。Taking Nan'an as an example, we summed up the characteristics of the unstructured Chinese address with the existing data, optimized the structure of existing address elements to meet the needs of address database construction and geocoding. And then, we proposed the Chinese semantic model construction method. In order to avoid the Chinese natural language ambiguity effectively and improve the address resolution confidence, we built a topology relationship between address elements to increase the context binding and introduce the support vector machine(SVM). The result shows that building Chinese address model can improve the standard database of geocoding support ability.

关 键 词：支持向量机 地址模型 中文语义 重庆市 

分 类 号：P208[天文地球—地图制图学与地理信息工程]