一种新的1／f过程小波模型

MODEL OF 1/f PROCESSES BASED ON DYADIC WAVELETS

给出了一种新的二进小波1/f过程模型,从理论上证明了一类谱指数为H的近似1/f过程可通过一簇平稳随机过程产生.由于所提方法利用了1/f过程小波系数的相关性,因而有效地减少了合成1/f过程的谱误差.数值实验结果表明,新模型很好地改进了已有模型.

A new model of 1/f processes based on dyadic wavelets was proposed. It was theoretically shown that a near 1/f process with spectral parameter H can be generated by a collect of stationary processes. Since the proposed method utilized the correlation among wavelet coefficients of a 1/f process, it efficiently reduces the spectral errors of the synthesized 1/f processes. Numerical experimental results demonstrate that the proposed model desirably improves the previous models.