| 基于改进支持向量机的高端装备制造业供应商分类研究 |
| |
| 引用本文： | 李坤,石春生,郑作龙,王成刚.基于改进支持向量机的高端装备制造业供应商分类研究[J].运筹与管理,2018,27(3):41-49. |
| |
| 作者姓名： | 李坤 石春生 郑作龙 王成刚 |
| |
| 作者单位： | 1.哈尔滨工业大学 管理学院,黑龙江 哈尔滨 150001; 2.辽宁工程技术大学 营销管理学院,辽宁 葫芦岛 125105; 3.苏州科技大学 商学院,江苏 苏州 215009 |
| |
| 基金项目： | 国家自然科学基金项目(71272176) |
| |
| 摘 要： | 针对现有供应商分类方法应用于高端装备制造业供应商所存在的局限性,从相互依赖视角构建了高端装备制造业供应商分类指标体系,提出了基于改进支持向量机的高端装备制造业供应商分类模型。该模型根据供应商误分代价不同,设计代价敏感支持向量机分类器,利用粒子群算法优化分类器的参数,并采用概率输出方法对多个优化的二类分类器的结果进行组合以实现多类分类。实验结果表明,该模型提高了现有方法的分类效果,可以降低总体误分代价,有效识别出对高端装备制造企业具有重大影响的供应商,为高端装备制造企业实施供应商分类管理提供了依据。
|
| 关 键 词： | 供应商分类 相互依赖 支持向量机 代价敏感学习 粒子群算法 |
| 收稿时间： | 2017-05-02 |
Research on Supplier Classification in High-end Equipment Manufacturing Industry Based on Improved Support Vector Machine |
| |
| LI Kun,SHI Chun-sheng,ZHENG Zuo-long,WANG Cheng-gang.Research on Supplier Classification in High-end Equipment Manufacturing Industry Based on Improved Support Vector Machine[J].Operations Research and Management Science,2018,27(3):41-49. |
| |
| Authors: | LI Kun SHI Chun-sheng ZHENG Zuo-long WANG Cheng-gang |
| |
| Affiliation: | 1.School of Management, Harbin Institute of Technology, Harbin 150001, China; 2.School of Marketing Management, Liaoning Technical University, Huludao 125105, China; 3.School of Business, Suzhou University of Science and Technology, Suzhou 215009, China |
| |
| Abstract: | There are some limitations of existing supplier classification methods which target high-end equipment manufacturing industry. From the perspective of interdependence, the paper divides suppliers of high-end equipment manufacturing industry into four types, including interdependence, supplier dominance, buyer dominance and independence. Meanwhile, the paper constructs a supplier classification index system based on literature analysis, interviews and expert judgment. The paper further proposes a classification model of high-end equipment manufacturing industry suppliers based on improved support vector machine. According to different misclassification costs of suppliers, the proposed model designs the cost sensitive support vector machine classifier. The particle swarm optimization algorithm is used to optimize model parameters, and the obtained optimized two-class models are assembled to realize multi-class classification according to probability outputs. The experimental results show that the proposed model can improve the classification performance of existing methods and reduce overall cost of errors. The proposed model also can identify suppliers which have significant impact on high-end equipment manufacturing enterprises effectively. The paper aims to provide a theoretical basis and practical guideline for high-end equipment manufacturing enterprises to implement supplier classification management. |
| |
| Keywords: | supplier classification interdependence support vector machine cost sensitive learning particle swarm optimization |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《运筹与管理》浏览原始摘要信息 |
|
点击此处可从《运筹与管理》下载全文 |
|
