A0193
Title: Thresholding-based robust estimation for generalized mixture models
Authors: Zhen Zeng - Nanjing University of Finance and Economics (China) [presenting]
Abstract: Finite mixture regression models are versatile tools for analyzing mixed regression relationships within clustered and heterogeneous populations. However, the classical normal mixture model often falls short when dealing with nonnormal or nonlinear regression data, especially in the presence of severe outliers. To address this, a novel generalized robust mixture regression procedure is introduced within the finite mixture regression framework. This procedure features sparse, scale dependent mean shift parameters, facilitating outlier detection and ensuring robust parameter estimation. The approach incorporates three key innovations:(1) A penalized likelihood approach using a combination of L0 (zero norm) and L2 (ridge) regularization to induce sparsity among mean shift parameters. (2) A close connection to the method of trimming, including explicit outlyingness parameters for all samples, simplifies computation, aids theoretical analysis, and eliminates the need for parameter tuning. (3) High scalability, allowing the implementation to handle nonnormal or nonlinear regression data. A threshold-based generalized expectation maximization algorithm has been developed to ensure stable and efficient computation. Simulation studies and real-world data applications demonstrate the effectiveness of this robust estimation procedure.
