摘要
探讨基于粗糙集的分类规则挖掘技术,提出一种融合不同数据源中的分类规则的方法.该方法能保证规则的完整性,即获得适用于全局的所有分类规则,并通过实验进行验证,结果表明该融合方法与直接在全局数据中挖掘分类规则的方法相比,具有运算量小、效率高的优势.针对大规模数据在分类求解中出现的过融合问题,应用剪枝策略进行实验,实验结果表明该剪枝策略正确可行,可以提高分类效果.
The mining technology of classification rules was approached based on rough sets and a method was proposed for amalgamating classification rules from multiple data sources. By using this method the completeness of rules could be guaranteed so that complete classification rules were obtained for overall situations. It was shown by the result of experimental verification of the given method that its computation labor was small and efficiency high. Finally, a pruning policy was tested for its application in dealing with the problem of over-amalgation occurred during the solutionof large-scale data classification.
出处
《兰州理工大学学报》
CAS
北大核心
2008年第1期87-90,共4页
Journal of Lanzhou University of Technology
关键词
粗糙集
多数据源
分类规则
融合
rough set
multiple data sources
classification rule
amalgamation