铁水硅含量的混沌粒子群支持向量机预报方法

提出一种基于混沌粒子群优化(CPSO)的支持向量回归机(SVR)参数优化算法, 并使用该算法建立高炉铁水硅含量预测模型(CPSO–SVR), 对某大型钢铁厂高炉铁水硅含量的实际采集数据进行预测, 结果表明基于混沌粒子群优化算法寻优的参数建立的铁水硅含量支持向量回归预测模型具有良好的预测效果. 与最小二乘支持向量回归机(LS–SVR)、使用粒子群优化算法训练的神经网络(PSO–NN)进行比较, CPSO–SVR模型对铁水硅含量进行预测时预测绝对误差小于0.03的样本数占总测试样本数的百分比达到90%以上, 预测效果明显优于PSO–NN, 且比LS–SVR稳定性更强, 可用于高炉铁水硅含量的实际预测, 表明混沌粒子群优化算法是选取SVR参数的有效方法.

The support vector regression based on the chaos particle swarm optimization algorithm for the prediction of silicon content in hot metal

An optimal selection approach of support vector regression parameters is proposed based on the chaos particle swarm optimization(CPSO) algorithm; A model based on the support vector regression to predict the silicon content in hot metal is established; and the optimal parameters of which is searched by CPSO. The data of the model are also collected from the No.3 BF in Panzhihua Iron and Steel Group Co. The results show that the proposed prediction model has better prediction results than the neural network trained by particle swarm optimization and least squares support vector regression algorithm; the percentage of samples with absolute prediction error less than 0.03 is higher than 90%, when predicting the silicon content by the proposed model. This indicates that the prediction precision can meet the requirement of practical production and demonstrates that the CPSO is an effective approach for parameter optimization of support-vector regression.