地理信息服务朴素贝叶斯分类及类别匹配研究  

Research of geographical information service-Nave Bayes classification and classification matching

作　　者：葛文[1] 成毅[1] 孙亚飞[2] 

机构地区：[1]信息工程大学测绘学院 [2]71834部队

出　　处：《测绘工程》2013年第2期5-9,共5页Engineering of Surveying and Mapping

基　　金：国家863计划资助项目(2009AA12Z228)

摘　　要：为进一步提高地理信息服务发现效率,文中提出地理信息服务朴素贝叶斯分类方法,并研究基于分类的地理信息服务匹配。地理信息服务朴素贝叶斯分类过程分为训练阶段和测试阶段,训练阶段主要是构造朴素贝叶斯分类模型,测试阶段主要是运用分类模型对地理信息服务进行分类测试。地理信息服务类别匹配基于本体实现,且只取完全匹配EXACT和包含匹配SUBSUME 2种精确匹配结果。实验结果表明,地理信息服务朴素贝叶斯分类的准确率可以达到78.1%,应用服务分类后,地理信息服务发现的查准率和查询时间均有较大改善。In order to improve the efficiency of geographical information service discovery, Naive Bayes classification method of geographical information service is proposed, and service classification matching is provided. The process of the classification includes training stage and test stage. Training stage aims to foim a Naive Bayes classification model, and test stage to make service classification tests by the model. Geographical information service classification matchmaking is based on the ontologies, and the results only confine to two precise values, EXACT and SUBSUME. Experiment results show that the accuracy of Naive Bayes classification can reach, to 78.1%, and the precision and inquiry time of geographical information service discovery are improved.

关 键 词：地理信息服务 朴素贝叶斯分类 服务类别匹配 服务发现性能 

分 类 号：P283[天文地球—地图制图学与地理信息工程]