论文部分内容阅读
提出一种基于独立成分分析(ICA)的最小二乘支持向量机(LS-SVM),用于时间序列的多步超前独立预测.用ICA估计预测变量中的独立成分(IC),用不含噪声的IC重新构建时间序列.利用-最近邻法(-NN)减小训练集的规模,提出一种新的距离函数以降低LS-SVM训练过程的计算复杂度,并用约束条件对预测值进行后处理.使用基于ICA的LS-SVM、普通LS-SVM与反向传播神经网络(BP-ANN),对多个时间序列进行对比预测实验.实验结果表明,基于ICA的LS-SVM的预测性能优于普通LS-SVM和BP-ANN.
A least-squares support vector machine (LS-SVM) based on independent component analysis (ICA) is proposed for independent prediction of time series with multi-step, ICA is used to estimate the independent component (IC) Noise IC to reconstruct the time series.A new distance function is proposed to reduce the training set size by using the nearest neighbor method (-NN) to reduce the computational complexity of the LS-SVM training process and to use the constraint conditions to predict the value SVM and Backpropagation Neural Network (BP-ANN) based on ICA are used to compare the performance of multiple time series prediction experiments.The experimental results show that the predictive performance of ICA-based LS-SVM Better than normal LS-SVM and BP-ANN.