基于邻域粗糙集优化支持向量机的备件分类研究  

Research on classification of spare parts based on Neighborhood Rough Set-Support Vector Machine

作　　者：杨华强[1] 尹亮[1] 赵青雨 夏唐斌[2] 郑美妹 YANG Huaqiang;YIN Liang;ZHAO Qingyu;XIA Tangbin;ZHENG Meimei(Hubei China Tobacco Industrial Co.,Ltd.,Wuhan 430040;School of Mechanical Engineering,Shanghai Jiao Tong University,Shanghai 200240)

机构地区：[1]湖北中烟工业有限责任公司,湖北武汉430040 [2]上海交通大学机械与动力工程学院,上海200240

出　　处：《机械设计》2023年第12期66-72,共7页Journal of Machine Design

摘　　要：针对现有备件分类中存在的备件种类繁多、属性复杂多样及分类标注不统一等问题,文中提出了一种基于邻域粗糙集的支持向量机(NRS-SVM)的多准则备件分类方法。首先,基于历史数据使用邻域粗糙集理论对备件属性进行约简,再将约简后的属性及数据输入支持向量机算法训练分类模型,最后可以将训练好的模型对真实的备件集进行分类。该方法对一家卷烟厂的实际备件数据进行试验验证,结果表明:基于邻域粗糙集的支持向量机在Z企业备件分类中具有高的分类准确率和优秀的泛化能力,验证了所提方法的有效性和优越性,从而更好地支持备件的管理。Currently,since the classification of spare parts suffer many problems,such as wide variety of spare parts,complex and diverse attributes,and inconsistent annotation of classification,in this article efforts are made to propose a multi-criteria method for classification of spare parts,which is based on Neighborhood Rough Set-Support Vector Machine(NRS-SVM).Firstly,the attributes of spare parts are simplified by the theory of neighborhood rough set based on the historical data;then,the simplified attributes and data are fed into the support vector machine's algorithm,in order to train the classification model;finally,the trained model is used to classify the real spare-part set.The method is used to verify the actual data on a cigarette factory's spare parts,and the results show that the support vector machine based on the neighborhood rough set has high accuracy in classification and better generalization ability in the classification of the Z company's spare parts.It is verified that this method is effective and superior to support the management of spare parts.

关 键 词：邻域粗糙集 支持向量机 多准则分类 备件分类 

分 类 号：F423[经济管理—产业经济]