贝叶斯框架下的LSSVM在热连轧机轧制力预报中的应用

钢铁研究学报, 2011, 23(6): 5-8.

贝叶斯框架下的LSSVM在热连轧机轧制力预报中的应用

刘俊，杨荃，何安瑞，解华华

北京科技大学高效轧制国家工程研究中心

基金项目: 2009年国家高技术研究发展计划“炼钢-连铸-轧制全流程仿真与优化控制设计平台”（2009AA04Z163）国家科技攻关课题

关键词：最小二乘支持向量机 , Bayesian evidence framework , prediction of rolling load , hot continuous rolling

Application of LSSVM With the Bayesian Evidence Framework to Prediction of Rolling Load of Hot Strip Rolling Mill

LIU Jun

Keywords： least square support vector machine , 贝叶斯推断 , 轧制力预报 , 热连轧

利用贝叶斯置信框架推断最小二乘支持向量机模型参数并建立贝叶斯最小二乘支持向量机非线性预测模型。在推断第1层确定模型最优参数，第2层确定正则化参数，第3层确定核参数。将该模型用于某1800热连轧轧制力的预测，在预测精度和速度上都取得了较好的效果。

Bayesian evidence framework was applied to least square support vector machines in order to optimize parameters and establish nonlinear prediction model. On the first level of inference, optimized model parameters were selected and on the second the regularization parameters were determined. The kernel parameters were tuned on the third level framework. The model was applied to the prediction of rolling load of 1800 hot strip mill and the simulation result indicated both the prediction accuracy and speed were better.

参考文献

上一张下一张

计量

下载量()
访问量()

作者相关

文章评分

您的评分：
1

0%
2

0%
3

0%
4

0%
5

0%

材料期刊网