Robust High-Dimensional Modeling with the Contaminated Gaussian Distribution

The contaminated Gaussian distribution represents a simple robust elliptical generalization of the Gaussian distribution; differently from the often-considered $t$-distribution, it also allows for automatic detection of outliers, spurious points, or noise (collectively referred to as bad points herein). Starting from this distribution, we propose the contaminated Gaussian factor analysis model as a method for robust data reduction and detection of bad points in high-dimensions. A mixture of contaminated Gaussian factor analyzers model follows therefrom, and extends the recently proposed mixtures of contaminated Gaussian distributions to high-dimensional data, i.e., where $p$ (number of dimensions) is large relative to $n$ (sample size). The number of free parameters is controlled through the dimension of the latent factor space. For each discussed model, we outline a variant of the classical expectation-maximization algorithm for parameter estimation. Various implementation issues are discussed, and we use real data for illustration.