FEMDA: a unified framework for discriminant analysis
This addresses robustness issues in discriminant analysis for data analysts, though it appears incremental as it builds on classical methods with a more flexible distributional assumption.
The paper tackles the problem of linear and quadratic discriminant analysis struggling with non-Gaussian or contaminated data by proposing a new framework based on arbitrary Elliptically Symmetrical distributions, resulting in a method that is simple, efficient, and robust compared to state-of-the-art approaches.
Although linear and quadratic discriminant analysis are widely recognized classical methods, they can encounter significant challenges when dealing with non-Gaussian distributions or contaminated datasets. This is primarily due to their reliance on the Gaussian assumption, which lacks robustness. We first explain and review the classical methods to address this limitation and then present a novel approach that overcomes these issues. In this new approach, the model considered is an arbitrary Elliptically Symmetrical (ES) distribution per cluster with its own arbitrary scale parameter. This flexible model allows for potentially diverse and independent samples that may not follow identical distributions. By deriving a new decision rule, we demonstrate that maximum-likelihood parameter estimation and classification are simple, efficient, and robust compared to state-of-the-art methods.