多级索引驱动的地名信息检索方法

Multilevel index-driven place name information retrieval method

针对地理空间数据交换和共享平台的地名数据日益增多及不同部门的数据差异所造成的地名检索效率低下问题,该文分析了平台中地名的表达特征,设计了面向地名信息的多级索引库组织方式,提出了地名特征词典的构建方法,设计并开发原型系统,实现了基于Lucene和地名特征词的检索框架。实验表明:多级索引通过基础索引、特征索引、分类索引三者联动的方式降低了地名检索的复杂度,具有较高的检索效率和准确度,应用于浙江省地理空间数据交换和共享平台取得了良好的效果。

Aiming at the low efficiency problem brought by increasing data amount of place name and differences between different data sources on the platform for geospatial data exchange and sharing. Ex- pression characteristics of place names were analyzed and place name information-oriented grouping method of multilevel index libraries was designed, the approach of building the characteristics dictionary was put forward further and a prototypical system was developed to implement the feature words and Lucene-based geographical name retrieval framework. Multilevel index managed the index of data source, the index of data characteristic and the index of classification. Experiments indicated that the proposed approach based on the combination of the basic index, feature index and classification index reduced the complexity of the place name retrieval and was of high indexing efficiency and accuracy. The deployment of this approach to the geospatial data exchange platform of Zhejiang Province has shown desired performance.