论文部分内容阅读
变尺度概率净化法是一种混沌时间序列非线性动力学降噪方法,该方法需对整个相空间点列做联合处理,因此算法的计算量和所需内存量会随着每个轨道点的修正参考点数和嵌入维数的增加而呈指数增长.根据变尺度概率净化法的特点,对前向概率和转移概率的估计方法作了一些改进,使算法的运算量减小到了原来的0·27左右,而降噪性能并没有下降,并提出了数据较长情况下的算法实现结构,大大降低了算法运行所需内存.
Variable scale probabilistic decontamination method is a chaotic time series nonlinear dynamics de-noising method, which requires joint processing of the entire phase space point sequence, so the amount of computation and the amount of memory required by the algorithm will vary with each track point The number of reference points and the number of embedding dimension is increased exponentially.According to the characteristics of the variable probability probabilistic purifying method, the methods of estimating the forward probabilities and transition probabilities are improved and the computational complexity is reduced to 0 · 27, while the noise reduction performance does not decline, and put forward the data to achieve a long time under the algorithm structure, greatly reducing the memory required for algorithm operation.