论文部分内容阅读
针对混沌时间序列的预测问题,考虑到单一核函数的最小二乘支持向量机无法明显提高预测精度,提出了一种组合核函数的最小二乘支持向量机预测模型,模型中采用多项式函数与径向基函数组合构建核函数.同时,还对遗传算法进行了改进,使之具有更快的收敛速度和更高的精度,改进的遗传算法适用于解决预测模型中的参数优化问题.通过典型的Lorenz时间序列、Mackey-Glass时间序列、太阳黑子数时间序列以及具有混沌特性的网络流量时间序列对该模型进行了验证.仿真结果表明所提出的模型是有效的.
Aiming at the prediction problem of chaotic time series, considering that the least-square support vector machine (SVM) with single kernel can not improve the prediction accuracy significantly, a least squares support vector machine (SVM) prediction model based on kernel function is proposed. The model uses polynomial function and path The kernel function is constructed based on the combination of basis functions.At the same time, the genetic algorithm is improved to make it faster convergence rate and higher precision.The improved genetic algorithm is suitable for solving the parameter optimization problem in the prediction model.Through the typical Lorenz time series, Mackey-Glass time series, sunspot time series and chaotic network traffic time series. The simulation results show that the proposed model is effective.