检索应用中地理信息资源相似度计算方法研究

On the method of similarity computation of geographic information resources in the retrieval

在地理信息资源检索应用中,资源属性间相似关系的计算还不够深入,地理信息资源的精准推荐及检索可视化应用尚难以实现。针对该问题,统一采用"相似度"对资源属性间的相似关系进行表达。资源的属性类型大部分为文本,其中具有语义差异,文中提出复合文本的相似度算法对文本类型属性进行匹配,并给出数值类型属性的"相似度"映射方案。在此基础上,实现资源时间、空间属性的相似度计算。基于所提出的相似度算法进行地理信息资源检索实验,实现地理信息资源属性间相似关系的量化,促进信息可视化在地理信息资源检索中的应用。

In the retrieval of geographic information resources,as the computation of relationship between the attributes of resources is not deep enough,it is difficult to recommend the geographic information resources accurately and display the results in a visual form.In view of the above problems,this paper uses relationship to express the similarity between resource attributes.The mostly-used type of attributes is text,which has much semantic difference.So a similarity algorithm is proposed to match the text attributes,and to obtain the similarity mapping scheme of numerical attributes.On this basis the similarity of time and space attributes are calculated.Based on the proposed similarity algorithm,geographic information resources retrieval experiments are carried out,which will realize the quantification of resource similarity relation and application of information visualization in geographic information retrieval.