【摘要】 针对支持向量机(SVM)的参数选择问题,提出了一种改进的PSO(QDPSO)算法。该算法采用同时对三个参数寻优的策略,克服了对每个参数单独寻优的弊端,可实现对SVM参数的精确、稳定、快速优化选择。给出了应用该方法的具体步骤,通过仿真实验证明该算法的有效性。通过将该方法得到的参数应用于SVM建模,得到了有机溶剂回收脱附过程的软测量模型。仿真结果表明,预测效果良好。更多还原

【Abstract】 Aiming at the parameters selection of support vector machines(SVM),an improved particle swarm optimiza-tion algorithm(quantum delta particle swarm optimization)was put forward.The algorithm adopted a strategy which op-timized three parameters synchronously.The disadvantage of optimizing for every parameter was avoided.It realized a accurate,stable and rapid optimal parameters selection of SVM.The material step of the method was shown.The simulated experiment proved the effectiveness of the algorithm above.A soft-sensor model of desorption-process about organic solvent recovery was established by applying the selected optimal parameters to modeling of SVM.The simulating result showed that the effectiveness of forecasting is satisfied.更多还原