Parameter Resolution of the Estimation Methods for Power Law Indices

Figure 1

Parameter resolution can be used in different models and algorithms, even for non-Gaussian noises. (a) The relationship between relative error and with different distinguishing method of linear model. The estimation algorithm is the least-squares method. (b) The relative error under different power law noise of the linear model. The estimation algorithm is the least-squares method. (c) The relative error under different SNR of the exponential model. The estimation algorithm is the nonlinear least-squares method. (d) The relative error under different power law noise of the exponential model. The estimation algorithm is the nonlinear least-squares method.

(a)

(b)

(c)

(d)