Clustering, classification, discriminant analysis, and dimension reduction via generalized hyperbolic mixtures

摘要: A method for dimension reduction with clustering, classification, or discriminant analysis is introduced. This mixture model-based approach based on fitting generalized hyperbolic mixtures a reduced subspace within the paradigm of analysis. data derived by considering extent to which group means and covariances vary. The members arise through linear combinations original data, are ordered importance via associated eigenvalues. observations can be projected onto subspace, resulting in set variables that captures most clustering information available. use gives robust framework capable dealing skewed clusters. Although increasingly demand across various application areas, many applications biological so some real examples sphere. Simulated also used illustration.