论文部分内容阅读
基于噪声的小波变换特点 ,结合小波包分解和模极大重构来抽取含噪信号的主分量 ,提出了一种基于最佳尺度分解和Volterra自适应滤波的分频预测算法 ,使用较少的模型训练样本 ,同时具有强的抗噪能力 .该算法克服了传统小波分解尺度选取的盲目性及单纯Volterra预测器抗噪性能的不足 ,数值仿真表明 ,针对含强噪声的非线性信号可进行有效预测 .
Based on the characteristics of wavelet transform of wavelet transform and wavelet packet decomposition and modular maxima reconstruction, the principal component of the noisy signal is extracted. A crossover algorithm based on best scale decomposition and Volterra adaptive filtering is proposed. Model training samples at the same time have strong anti-noise ability.The algorithm overcomes the blindness of the traditional wavelet decomposition scale selection and the shortcomings of simple Volterra predictor anti-noise performance, numerical simulation shows that for non-linear signals containing strong noise can be effective Forecast.