Fast Parameter Estimation Method for Generalized Gaussian Distribution

In signal processing and communication fields, estimating parameters of Generalized Gaussian Distribution (GGD) represents a crucial task. Various methods have been developed to achieve more accurate data fitting. Maximum Likelihood Estimation (MLE) stands as one rapid approach for GGD parameter estimation, though it requires numerical optimization and incurs significant computational costs. To address this limitation, a local extrema-based estimation method has been proposed. This technique leverages GGD properties by estimating parameters within individual local regions and subsequently aggregating results for final parameter determination. The implementation typically involves: 1) detecting local maxima/minima in the data distribution, 2) applying moment-matching or quantile-based estimation within each segment, and 3) employing weighted averaging for final parameter consolidation. This method significantly reduces computational overhead while maintaining satisfactory estimation accuracy, making it a powerful tool for GGD parameter estimation. Key algorithmic advantages include parallelizable local computations and adaptive region sizing based on distribution characteristics.