摘要
提出了基于支持向量回归的钢材力学性能建模方法,采用遗传算法优化支持向量回归模型的参数,避免了参数选择的盲目性,使得支持向量回归模型的预测性能有了显著提高。将此方法应用于实际钢厂的钢材力学性能预报中,模型的训练与验证数据都来自于实际的过程,结果表明采用遗传优化的支持向量回归模型对钢材力学性能具有很好的预估性能。
The prediction of the Mechanical Property of the steels was discussed based on Support Vector Regression （SVR）, and genetic algorithms were introduced to optimize the parameters of SVR model, which could avoid the blindness when defining the parameters to improve the prediction capability greatly. The method is applied to the Prediction of Mechanical Property of Steel Materials, and training data of the model are all based on the actual process. Result shows that SVR model optimized by genetic algorithms is highly beneficial to the estimate of Mechanical Property of Steel Materials.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2009年第4期1192-1194,共3页
Journal of System Simulation
基金
北京市教育委员会重点学科共建项目资助(XK100080537)
关键词
力学性能
支持向量回归
遗传算法
参数优化
mechanical property
support vector regression
genetic algorithm
parameters optimization
