A novel estimator for the shape parameter of the generalized Gaussian distribution (GGD) is proposed based on the method of logarithmic cumulants. First, the expression of the log-cumulant for the… Click to show full abstract
A novel estimator for the shape parameter of the generalized Gaussian distribution (GGD) is proposed based on the method of logarithmic cumulants. First, the expression of the log-cumulant for the GGD is theoretically derived. As a result, a simple equation for the estimation of the shape parameter is obtained. The processing procedure of the new estimator for practical applications is also provided. Numerical experiments are used to verify the excellent estimation performances of the proposed estimator for both large and small samples. Moreover, experiments using a measured CARABAS-II data set also validate the superiority of the proposed estimator.
               
Click one of the above tabs to view related content.