Abstract: This paper proposes a novel method for data source management in a grid-based information integration system for huge volume of dynamic data. It includes an object-oriented metadata description specification, a data query mechanism and a schema integration strategy，, which can effectively solve the management problem of dynamic data sources in data gird. An effective class mining algorithm based on clustering technique is also used to optimize the performance.

Key words: data fusion, grid computing, information integration

摘要： 针对网格海量动态数据高效融合的要求，提出提出一个套新的数据源管理方案，主要包括一套一套面向对象的元数据描述规范和、基于此规范的数据查询机制以及、模式融合策略，以解决传统网格中动态异构数据源难以维护的问题。给出，一种基于聚类的有效类挖掘算法。实验证明，该能有效地解决传统网格中动态异构数据源难以维护的问题。另外提出一种基于聚类的有效类挖掘算法，实验证明算法能大大优化系统的性能。

关键词: 数据融合, 网格计算, 信息集成

CLC Number: 

• TP311