信息集成中基于描述器的对象匹配

Describer-relied object matching in information integration

提出基于描述器的对象匹配(DROM)方法,通过增加一个私有属性关联层来扩展已有的方案,从而提高匹配正确率。探讨对象匹配环境中的知识重用,包括知识类型的重用和任务无关分类器的重用。实验结果表明,增加描述器能使DROM获得更多的匹配知识,从而可以提高DROM的匹配正确率。

Describer-relied object matching （DROM） is proposed, which extends previous solutions by adding a layer that correlates private attributes to enhance matching accuracy. Knowledge reuse in the object-matching context, including knowledge type reuse and task-irrelative classifier reuse is discussed. The experimental result shows that adding more describers can make DROM access more matching knowledge, therefore the accuracy of DROM might be enhanced.