改进的粗糙集属性重要度

发布时间：2018-06-12 20:37

本文选题：粗糙集 + 属性重要度　；参考：《计算机工程与应用》2017年03期

【摘要】：粗糙集理论知识库的属性重要度,体现的是去掉某个或某些属性前后的知识库分类变化的程度。对现有粗糙集理论的属性重要度确立方法的不足,充分考虑条件属性对决策的直接和间接的影响,提出一种新的基于粗糙集属性依赖度的属性重要度确定方法。此外,针对原有属性重要度与改进重要度的差别,讨论改进的属性重要度的意义,并证明改进的属性重要度更加可信。最后,利用改进的方法对机械故障属性重要度进行仿真;对比原有属性重要度的数据,改进方法获得的数据不但更符合属性约简结果,并且具有更大区分度,十分有利于决策者快速做出判断。
[Abstract]:The attribute importance degree of rough set theory knowledge base reflects the degree of classification change of knowledge base before and after removing one or some attributes. Considering the direct and indirect influence of conditional attributes on decision making, a new method of attribute importance determination based on attribute dependency of rough sets is proposed. In addition, in view of the difference between the original attribute importance and the improved importance, the significance of the improved attribute importance is discussed, and it is proved that the improved attribute importance is more reliable. Finally, the improved method is used to simulate the importance of mechanical fault attribute, and compared with the data of original attribute importance, the data obtained by the improved method not only accord with the result of attribute reduction, but also have a greater degree of distinction. It is very helpful for decision makers to make quick judgments.
【作者单位】：广东石油化工学院理学院;
【基金】：国家自然科学基金(No.11426070) 广东省青年创新人才项目(No.2014KQNCX202) 广东省普通高校特色创新项目(No.2014KTSCX167)
【分类号】：TP18

【相似文献】