基于D-S证据理论的XML文档潜在信息获取算法

传统的XML文档检索方法主要是基于关键词匹配的检索,忽略了关键词的语义信息和蕴涵于信息组合中的潜在信息。针对上述问题,提出了基于D-S证据理论的XML文档潜在信息的获取算法。该算法通过引入本体定义了概念间的语义关系和信息的组合方式,提出了基于D-S证据理论的检索模型和指标权重的计算方法,并结合似然函数设计了一个动态的阈值,有效地消除语义匹配过程中存在的不确定性,解决了信息组合中潜在信息的获取问题。此外,还将该算法应用于电子政务领域个人和企业敏感信息的检测中,实验证明了该算法比传统的方法有着更高的查准率和查全率。

XML document latent information extractionalgorithm based on D-S evidence theory

Traditional XML document retrieval methods are mainly based on keywords' match, which ignore keywords' semantics and latent information contained in information combination. This paper proposed an algorithm of XML document latent information extraction based on D-S evidence theory. Firstly it used ontology to define the relationships between semantic concepts and the combination mode, and next proposed a retrieval model based on D-S evidence theory. Then it presented the computation of evidence weight, and finally designed a dynamic threshold with plausible function. It solved the problems of uncertainty in semantic match and retrieve of latent information. Furthermore, it presented the algorithm's application in the detection of personal and enterprises' sensitive information in e-government domain. The experiment proves that the proposed algorithm has higher precision and recall.