欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
针对最小二乘支持向量机（LS-SVM）的参数寻优问题,提出一种基于混合量子粒子群算法（HQPSO）的LS-SVM参数选择方法,以提高LS-SVM模型的学习性能和泛化能力。该算法结合QPSO算法的全局优化能力和Powell的局部寻优能力,分别对粒子初始位置、新局部最优位置以及全局最优位置进行Powell局部寻优,提高求解速度和解的精确性。利用测试函数对该建模方法进行仿真测试,与PSO LS-SVM模型进行比较,并利用湿法炼锌净化过程现场数据进行工业验证。研究结果表明：HQPSO LS-SVM模型具有较好的泛化性能,模型预测精度高,预测结果满足工艺生产的要求。
英文摘要：
Aiming at the parameter optimization problem in least squares support vector machine,a hybrid QPSO algorithm for LS-SVM parameter selection was proposed to improve the learning performance and generalization ability of LS-SVM model.The Powell algorithm was used to obtain the initial position,local optimal position and global optimal solution.The proposed hybrid QPSO method combines with the global search ability of QPSO algorithm with the local search ability of Powell algorithm,to improve the solving speed and the accuracy of the solution.This modeling method was validated by test function firstly and compared with the PSO LS-SVM model.Then the production data from a purification process of zinc hydrometallurgy was used to test the model precision.The test results show that the proposed model has better generalization performance and higher precision.
同期刊论文项目
同项目期刊论文
Nonparametric density estimation of froth colour texture distribution for monitoring sulphur flotati
Kinetic Modeling and Parameter Estimation for Competing Reactions in Copper Removal Process from Zin
Operational patterns recognition using multi-class cost-sensitive learning for alumina evaporation p
Recognition of the operational statuses of reagent addition using dynamic bubble size distribution i
期刊信息
位置：