| 最小残差法加速局部加权LSSVM求解及其应用 |
| |
| 引用本文： | 林超,;王华杰.最小残差法加速局部加权LSSVM求解及其应用[J].微型电脑应用,2014(11):8-11. |
| |
| 作者姓名： | 林超 ;王华杰 |
| |
| 作者单位： | [1]中国石油大学(华东),网络及教育技术中心,青岛266580; [2]山东财经大学管理科学与工程学院,济南250014 |
| |
| 基金项目： | 国家自然科学基金（No.11326203）;山东省自然科学基金（No.ZR2013FQ034） |
| |
| 摘 要： | 局部加权最小二乘支持向量机回归模型（LocalWeighted Least Squares Support Vector Machines,LW-LSSVM）是一种在线学习模型,该类模型需要根据训练样本权重的调整不断重新进行训练.高效稳定的学习算法是LW-LSSVM模型取得成功应用的关键.分别采用最小残差法（MINRE）、共轭梯度法（CG）、零空间法和Cholesky分解算法求解WL-LSSVM模型.基准数据库上的数值实验表明最小残差法的计算时间最短,具有良好的数值稳定性.随后,应用基于MINRES的WL-LSSVM建立了高炉铁水硅含量的在线预测模型,仿真实验表明与LSSVM相比LW-LSSVM模型具有更高的预报精度和自适应性.
|
| 关 键 词： | LSSVM 局部加权 最小残差法 铁水硅含量 |
Minimum Residual Method to Accelerate the Local Weighted LSSVM and its Application in Industry |
| |
| Affiliation: | Lin Chao, Wang Huajie (1. Network and Educational Technology Center, China University of Petroleum, Qingdao 266580, China; 2. School of Management Science and Engineering, Shandong University of Finance and Economics, Jinan 250014, China) |
| |
| Abstract: | Local weighted least square support vector machines regression model is a kind of e-learning model,the model needs constantly retraining according to the adjustment of the training sample.The stable and efficient learning algorithm is the key to the successful application of LW-LSSVM model.This paper utilized minimum residual method （MINRE）,conjugate gradient （CG）,null space method and Cholesky decomposition algorithm respectively to solve WL-LSSVM model.The numerical experiment of benchmark database indicates that the computing time of minimum residual method is the minimum as well as good numerical stability.Afterwards WL-LSSVM based on MINRES is used to establish the e-learning prediction model of the silicon content in hot metal.The simulation experiment shows that compared with LSSVM,LW-LSSVM has higher prediction precision and adaptability. |
| |
| Keywords: | LSSVM Local Weighted MINRE Silicon Content |
| 本文献已被 维普 等数据库收录！ |
|
