摘要
传统控制图通过判异准则分析生产过程状态,这一方法无法识别过程质量的具体异常原因。提出了一种基于自适应增强集成算法(AdaBoost)和支持向量机(SVM)结合的控制图异常模式识别方法。仿真结果表明,相较于单一的分类器和集成分类器,改进的集成支持向量机对控制图异常模式的识别率更高。最后将所提出的模型用到物流港口配煤的模拟实验中,结果表明,基于AdaBoost和SVM结合的控制图异常模式识别方法具有较高的检测效率。
The traditional control chart analyzes the state of the production process based on the out-of-control(OOC) criterion, which cannot identify the specific cause of the quality anomaly of the process. In this paper, we proposed a control chart anomaly pattern recognition method based on the combination of adaptive boost ensemble algorithm(AdaBoost) and support vector machine(SVM). The simulation result showed that, compared with the single classifier and the ensemble classifier, the improved ensemble support vector machine has a higher recognition rate for the anomaly pattern of the control chart. Finally, we applied the proposed model in the simulation experiment of the coal blending operation at a logistics terminal, showing that the control chart anomaly pattern recognition method based on the combination of AdaBoost and SVM has high detection efficiency.
作者
张莹
褚娜
ZHANG Ying;CHU Na(School of Logistics Engineering,Wuhan University of Technology,Wuhan 430063;Engineering Research Center of Port Logistics Technology&Equipment,Ministry of Education,Wuhan 430063,China)
出处
《物流技术》
2022年第6期54-59,共6页
Logistics Technology
关键词
控制图
集成算法
支持向量机
控制图模式识别
异常模式识别
港口配煤
control chart
ensemble algorithm
support vector machine
control chart pattern recognition
abnormal pattern recognition
port coal blending