A NOVEL MODELING METHOD BASED ON SUPPORT VECTOR DOMAIN DESCRIPTION AND LS-SVM FOR STEEL-MAKING PROCESS

摘要：

A novel modeling approach to predict the end-point phosphorus content in electric-arc furnace steel -making plant is proposed. The approach includes two procedures. Firstly, it detects the abnormal sample data point caused by disorder operating mode in the original training set with Support Vector Domain Description method and erases these abnormal samples; then, it reconstructs a new training set with these clean sample data. Secondly, the predictive model is obtained by using Least Square Support Vector Machines and the new training set. Through the comparative experiments between the proposed approach in this paper and the direct modeling approach using Least Square Support Vector Machines with the original training set, the results show that the proposed approach has superiority in the end-point phosphorus content predictive task in steel-making process.

关键词： Support vector domain description Least square SVM Complez industrial process End-point predictive model

作者: SHI-YU PENG GUO-YUN ZHANG HONG-MIN LI

作者单位: Dept.of Physics and Electronics Information, Hunan Institute of Science and Technology, Yueyang 4140 Dept.of Physics and Electronics Information, Hunan Institute of Science and Technology, Yueyang 4140 College of Electrical and Information Engineering, Hunan University, Changsha 410082, China

会议类型: 国际会议

会议名称: 2008 International Conference on Machine Learning and Cybernetics(2008机器学习与控制论国际会议)

会议地点: 昆明

会议语种:英文

页码: 2229-2233

在线出版日期: 2008-07-12（万方平台首次上网日期，不代表论文的发表时间）