摘要
针对属性约简时组合众多的情况,笔者提出了一种基于改进二元萤火虫算法和邻域粗糙集组成的属性约简方法。首先,通过改变移动策略、仿生启发和sigmoid函数映射改进成一种二元萤火虫算法,结合邻域粗糙集构建一种属性约简方法;然后,在UCI数据集上进行了实验验证;最后,在构建的数据集上进行应用。实验表明,与其他两种约简方法相比,该种属性约简方法提升了分类准确率,在民族团结进步创建评价中的应用也证明了其有效性。
In view of the fact that there are many combinations in attribute reduction,an attribute reduction method based on hybrid population binary firefly algorithm and neighborhood rough set is proposed.First of all,a binary firefly algorithm is improved by changing the mobile strategy,bionic heuristic and sigmoid function mapping;secondly,an attribute reduction method is constructed by using the algorithm as the search strategy and combining with the neighborhood rough set;then,it is verified by experiments on the UCI data set;finally,it is applied in the data set constructed by relevant indicators about creation of national unity and progress.Experiments show that compared with the other two reduction methods,this attribute reduction method improves the classification accuracy,and its application in the evaluation of national unity and progress has also proved its effectiveness.
作者
彭鹏
姜志梅
Peng Peng;Jiang Zhimei(Law School,North Minzu University,Yinchuan Ningxia 750021,China;Department of Economy and Trade,Ningxia Vocational and Technical College of Finance and Economics,Yinchuan Ningxia 750021,China)
出处
《信息与电脑》
2020年第12期4-6,共3页
Information & Computer
基金
北方民族大学中央高校基本科研业务费专项资金资助(项目编号:2018XYSFX02)。
关键词
邻域粗糙集
属性约简
民族团结进步创建
neighborhood rough set
attribute reduction
national unity and progress creation
