中文摘要：

顾及到地理领域语义相似度计算模型考虑因素过于单一、主观性较强等问题,针对本体模型的结构特点,提出一种计算节点密度的新方法,并从模型概念间的关系类型、节点密度、节点深度等方面分析本体概念相似度的计算,将其归并为距离因素.基于本体层次网络结构计算语义信息量,该方法不依赖于专家经验,具有客观性.结合语义距离、信息量、属性等影响相似度的因素,提出一种计算概念问语义相似度的综合算法,该算法考虑到不同的影响因子在语义相似度计算中的重要程度不同,从而赋予地理本体关系不同的权值.通过对土地利用分类中实体的语义相似度进行实例验证,表明提出的算法能有效改善语义相似度计算的准确性和有效性,能够获得更符合认知的信息检索结果.

英文摘要：

Because traditional semantic similarity algorithms of geographical domain have single impact factor,strong subjectivity and other defects,a weighted semantic similarity algorithm based on geograpical domain ontology was proposed here.The structural characteristics of an ontology model were taken into account,a new method for calculating the node density proposed,and ontological concept similarity analyzed from the aspects of the edge type between concepts,node density and depth,which were merged into the distance factor.Semantic information was calculated based on ontological hierarchical network,which does not rely on the expertise and possesses objectivity.Combining semantic distance,information and property,a integrated semantic similarity algorithm was proposed,which considered that different impact factors would have a different important degree in the calculation of semantic similarity and different weights were given to geographic ontology relations.Part of the hierarchical network of the geographical domain land use classification ontology was used in the experiment for analysis and comparison.The results showed the algorithm can effectively improve the accuracy and validity of semantic similarity calculation of the concepts,and more information retrieval results in line with people＇ s perception can be obtained.