基于本体概念相似度的遥感信息服务匹配研究

Study of remote sensing information service matching based on similarity of ontology concept

构建了一个遥感信息领域本体,基于领域本体和Word Net词典对遥感信息服务进行语义扩展,提出了一种基于本体概念相似度的遥感信息服务匹配方法,并对Leacock和Chodorow语义相似度计算模型进行改进。实验结果表明,该改进模型比距离模型和信息量模型都有提高,基于本体概念相似度的遥感信息服务匹配方法的查全率和查准率都能达到70%以上,较关键字匹配方法有显著提高。

It builds a remote sensing information domain ontology in this paper. Based on the semantic extended remote sensing information service by the domain ontology and Word Net dictionary, this paper proposes a remote sensing information service matching method based on ontology concept similarity. In this method, the Leacock and Chodorow model is improved based on information content theory. Experimental results show that the improved Leacock and Chodorow model is more accurate and efficient than the distance-based model and information content model. And the remote sensing information service matching method based on ontology concept similarity has improved significantly in accuracy and efficiency. Its recall ratio and precision can achieve 70% above, which is better than keyword matching method.