摘要
提出了两种高预测效率、高泛化能力的数值计算方法,一种方法是小波神经网络(Wavelet NeuralNetwork,WNN)算法;另一种是基于小波尺度函数的WSK-SV(Wavelet Scaling Kernel-Support Vector)算法.WNN算法将小波函数与BP神经网络方法相结合,通过输入层、隐含层、输出层间的连接权重以及隐含层使用的激励函数构成了这种算法的关键技术;WSK-SV算法将小波的尺度函数与SV(Support Vector)方法相结合,使这种算法既保持了SVM(Support Vector Machine)的优点,又具有很好的泛化能力.上述两种算法都属于计算智能(Computational Intelligence,简称CI)方法并用于人机系统的性能预测.
Two numerical computational methods with high forecast efficiency and high generalization ability are introduced,one being the Wavelet Neural Network（WNN） method and the other the Wavelet Scaling Kernel-Support Vector（WSK-SV） method.The WNN method combines the wavelet function and the BP Neural Network method.The key technique consists of link-weight of the input layer,the hidden layer,the output layer,and the excitation function of the hidden layer.On the other hand,the WSK-SV method does the scaling function and the Support Vector（SV） method with the advantage of Support Vector Machine（SVM） as well as good generalization ability.Both methods above are included in Computational Intelligence（CI） method.They are both applied in performance forecast of human-machine system.
出处
《青岛理工大学学报》
CAS
2012年第5期82-85,共4页
Journal of Qingdao University of Technology
