摘要
提出一种改进的离群点检测方法,采用参数更新的支持向量数据描述的离群点检测方法,并引入贝叶斯分类原理对离群点分类,对校正离群点后的数据用最小二乘支持向量机建模并预测。工业聚丙烯熔融指数软测量模型的应用结果表明:该方法预测精度更高,泛化能力更强。
A soft sensor research based on improved outlier detection is proposed and the method of outlier dotcction based on support vector data description （SVDD） is introduced in this paper. To determine whether the outliers get is authentic, the concept of outlier classification is put forward. The application of industrial polypropylene melt index soft measurement modeling has indicated that method has better prediction accuracy and generalization performance.
出处
《自动化与信息工程》
2015年第4期18-23,共6页
Automation & Information Engineering
关键词
软测量
离群点检测
贝叶斯分类
Soft Measurement
Outlier Detection
Bayesian Classification
