一种基于相似粗糙集技术的案例库维护

最近的研究工作突现了在案例推理过程中案例库维护的重要性,越来越多的人认为基于案例推理系统包含了案例库维护的有关过程(Review和Restore)。案例库维护作为CBR研究的一个分支,已经研究出不同的案例库维护策略,其中一些是限制案例库的规模,由此引发了CBR系统的能力与效率问题。相似粗糙集技术可以有效地利用差别矩阵,通过不同的相似度阈值发现以及处理案例库的冗余,有选择地删除多余的案例;同时案例库的覆盖度不降低,减少了案例适应性修改的代价,从而确保了CBR系统的能力与效率的兼顾。

A Similarity Rough Set Technique Approach to Case-base Maintenance

Recent work highlighted the importance of the maintenance in a Case- Based Reasoning (CBR)process proposing a new CBR cycle that includes some maintenance steps- Review and Restore.In recent years,a great amount of interest has been paid to Case- Base Maintenance(CBM),a CBR research branch,different strategies have been de- veloped.Some of them focus on developing methods that limit the number of cases and concerns about competence- preserving or efficiency- preserving.Based on similarity rough set technique,the paper proposes a case selection algorithm that can use Discernible Matrix efficiently,search and deal with the redundant information as well as choose suitable cases to be deleted. These will ensure a large lower bound on the coverage of the resultant case base,which is tied closely to adaptation costs,and can guarantee the two all,competence and efficiency while maintaining satisfactory similar accuracy in case base.