基于WA SVM模型的高炉铁水含硅量预测

基于小波在处理非线性、非平稳随机信号和支持向量机在解决非线性、高维数、小样本等问题的优点，提出了一种二者组合的预测模型。先用小波变换将铁水含硅量的时间序列分解成不同的高频和低频层次，对不同层次构建支持向量机模型进行预测，然后通过序列重构得到原始时间序列的预测结果。利用山东莱钢1号高炉在线采集的数据作为应用案例，WA SVM组合模型与工程常用的AR模型和单一的最小二乘支持向量机模型的预测结果比较，预测精度有明显提高。

Application of WA-SVM Combined Model to Predict Silicon Content in Hot Metal

Based on the fact that wavelet is suitable for processing nonlinear, non-stationary random signals and support vector machine excel at solving nonlinear, small-sample, high dimensional problems, the paper proposes a combined model of wavelet analysis （WA） and support vector machine （SVM）. It decomposes the time series of original silicon content in hot metal to different layers through wavelet analysis. Different SVMs are built to predict each layer, and finally to obtain the predicted results of the original time series by composition. Taking No. 1 blast furnace of Laiwu Iron and Steel Group Co. as an example, application result shows that the result of WA-SVM model is better than that of the AR model frequently applied in the project and the single least squares support vector machine and the prediction accuracy is elevated obviously.