Data‐Driven Wave Distribution Function Method Based on Generalization Error Minimization

Abstract Accurately estimating the propagation direction of plasma waves is essential for preserving space infrastructure, as it provides insight into the global spatial distribution of plasma in geospace. This study proposes an improved wave distribution function (WDF) estimation algorithm based on the conventional Markov random field (MRF) model. In the MRF‐based WDF method, the optimal WDF is obtained by minimizing an objective function that balances the discrepancy between theoretical and observed spectral matrices with the smoothness of the WDF. We introduce a data‐driven approach to automatically determine the weighting between these terms by minimizing the generalization error. Validation using computer‐generated data demonstrates that the proposed method yields high‐quality WDF estimates, as confirmed through visual evaluation.