arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1003.3082cs.AIcs.DB
收藏

P2P环境下基于模式和本体变化的协议维护

Agreement Maintenance Based on Schema and Ontology Change in P2P Environment

Gunadarma University(贡达玛大学)

L. Y. Banowosari, I. W. S. Wicaksana, A. B. Mutiara

摘要:

提出了一种基于语义距离的语义一致性维护方法,通过计算局部模式或本体的变化来实现。这种方法在动态自治环境中具有重要意义,目前的方法假设静态环境中的协议或映射。本研究的贡献在于提出一个基于语义协议维护的P2P环境下的框架。该框架基于两级混合P2P模型体系结构,由两种类型的节点组成:(1)超级节点,用于注册和管理其他节点;(2)简单节点,作为简单节点,输出和共享自己的内容。本文提出了一种P2P环境下的语义一致性维护模型,解决了现有的基于本体和局部模式的P2P语义一致性维护方法在静态和动态情况下存在差异,无法及时了解语义一致性变化的问题。主要问题是如何计算局部模式或公共本体的变化,并根据计算结果来确定哪种算法在维护协议。在印度尼西亚的工作匹配域的实验已经完成,以显示多远的性能的方法。实验结果表明:(i)修改越多,F-测度值越小;(ii)不同修改类型(添加、删除、重命名)的F-测度值没有显著差异;(iii)正确选择算法会提高F-测度值。

英文摘要:

This paper is concern about developing a semantic agreement maintenance method based on semantic distance by calculating the change of local schema or ontology. This approach is important in dynamic and autonomous environment, in which the current approach assumed that agreement or mapping in static environment. The contribution of this research is to develop a framework based on semantic agreement maintenance approach for P2P environment. This framework based on two level hybrid P2P model architecture, which consist of two peer type: (1) super peer that use to register and manage the other peers, and (2) simple peer, as a simple peer, it exports and shares its contents with others. This research develop a model to maintain the semantic agreement in P2P environment, so the current approach which does not have the mechanism to know the change, since it assumed that ontology and local schema are in the static condition, and it is different in dynamic condition. The main issues are how to calculate the change of local schema or common ontology and the calculation result is used to determine which algorithm in maintaining the agreement. The experiment on the job matching domain in Indonesia have been done to show how far the performance of the approach. From the experiment, the main result are (i) the more change so the F-measure value tend to be decreased, (ii) there is no significant different in F-measure value for various modification type (add, delete, rename), and (iii) the correct choice of algorithm would improve the F-measure value.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑