异构Web对象共指冲突识别与消解研究

Research on heterogeneous Web object coreference recognition and resolution algorithm

针对现有的语义Web中对象共指的消解工作研究不足,提出了一种异构Web数据流对象共指消解的技术.首先,定义了Web数据流对象的对象共指冲突识别与消解框架;然后,提出了一种基于描述逻辑的对象共指识别技术,归纳给出了描述逻辑的层次依赖关系并提出了基于依赖图的冲突遍历识别算法;并且,在冲突消解方面,定义了冲突的三种类型、冲突消解匹配模式与仲裁消解算法;最后,分别用测试集检验了本文提出的异构Web数据流冲突识别与消解技术的应用效果.

To resolve the coreference problem of the semantic Web data stream, this paper proposed aheterogeneous Web data stream object coreference resolution. First, the paper defined a framework torecognize and resolve Web data stream object coreference. Secondly, the paper described a conflict rec-ognition system structure based on rule inference. The conflict recognition algorithm makes use of a fastrule matching, which improves the conflict recognition process efficiency and reduces the total refers toobjects in space, time conflict and complexity of the digestion. In third, the paper defines three types ofconflict and arbitration resolution algorithms for conflict resolution. Lastly, test sets are used in this pa-per to examine the application of heterogeneous Web data stream conflict recognition and coreferencetechnology.