基于LS-SVM的烧结矿碱度预报研究  

Prediction of sinter basicity based on a least square support vector model

作　　者：宋强[1] 程国彪[2] 常卫兵[2] 李华[2] 

机构地区：[1]安阳工学院机械工程系,河南安阳455000 [2]安阳钢铁(集团)公司烧结厂,河南安阳455004

出　　处：《钢铁研究》2008年第6期10-13,共4页Research on Iron and Steel

基　　金：河南省教育厅自然科学基金(200612001)

摘　　要：开发了最小二乘支持向量机(LS-SVM)模型,并用于对烧结矿碱度进行预测。仿真结果证明,本模型能在小样本贫信息的条件下对烧结矿碱度做出比较准确的预测。此种模型具有预测精度高、所需样本少、计算简便等优点。和BP神经网络算法相比,最小二乘支持向量机算法有很好的应用前景和推广价值。A least square support vector machine （LS-SVM）model is developed and applied in the prediction ot sinter basicity. The results of model calculation show that the sinter basicity can be accurately predicted in a condition of small amount of samples. It is concluded that the LS-SVM model has advantages of high precision, less samples required and simple calculation. Compared with BP neural network algorithm, the LS- SVM model has good prospects in practical application.

关 键 词：碱度 最小二乘支持向量机 预测 神经网络 

分 类 号：TF123.11[冶金工程—粉末冶金]