改进DM-SVDD算法的异常检测研究及应用  

Research and Application of Anormaly Detection Based on Improved DM-SVDD Algorithm

作　　者：王杰 张雪英 李凤莲 杜海文 于丽君 马秀 WANG Jie;ZHANG Xueying;LI Fenglian;DU Haiwen;YU Lijun;MA Xiu(College of Information and Computer, Taiyuan University of Technology, Taiyuan 030024, China;CETC New Energy Technology Co., Ltd, Taiyuan 030024, China)

机构地区：[1]太原理工大学信息与计算机学院,太原030024 [2]山西中电科新能源技术有限公司,太原030024

出　　处：《太原理工大学学报》2021年第5期764-768,共5页Journal of Taiyuan University of Technology

基　　金：山西省科技重大专项资助项目(20181102008);山西省研究生教育创新计划基金资助项目(2020SY522)。

摘　　要：针对传统异常检测模型在数据不平衡情况下对少数异常类样本识别效果较差的问题,提出了一种基于改进扩散映射的支持向量数据描述算法(DM-SVDD),构建新的模型并将其应用于工业异常检测。通过引入欧氏距离和马氏距离两种距离度量公式,构造新的近邻图改进扩散映射算法,结合支持向量数据描述算法进行建模,新模型不仅提高了对多数正常类样本的识别性能,且对少数异常类样本的检测性能也优于传统模型。实验数据选取多晶硅铸锭生产过程中的配料数据,研究结果表明:对于异常类样本较少所形成的不平衡数据,与传统的异常检测模型相比,所提出的改进模型可使G-Mean最优提升15.73%,F-Score最优提升19.37%,满足工业异常检测的需求,可用于指导实际生产过程,降低生产成本。Aiming at the problem that traditional anomaly detection model has poor recognition effect on a small number of abnormal samples under the condition of data imbalance,in this paper we proposed a support vector data description algorithm combined with improved diffusion maps(DM-SVDD),constructed a new model and applied it to industry abnormal detection.The diffusion mapping algorithm was improved by introducing Euclidean distance and Mahalanobis distance to construct a new neighbor graph.Combined with support vector data description algorithm for modeling,the new model improved the recognition performance of normal samples,and the detection performance of abnormal samples was better than that from traditional models.Experimental data were selected of polysilicon ingot data sets.The results show that for an unbalanced data set formed by fewer abnormal samples,compared with traditional anomaly detection model,the model proposed in this paper can increase G-Mean optimally by 15.73%and F-Score optimally by 19.37%,which meet the requirements of industrial anomaly detection.The model can be used to guide the actual production process and reduce production costs.

关 键 词：支持向量数据描述 扩散映射 异常检测 不平衡数据 多晶硅数据 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]