B2012
Title: Skewed normal classification in high-dimensional data
Authors: Haesong Choi - Florida state university (United States) [presenting]
Abstract: A considerable number of studies have been devoted to high-dimensional classification models under the assumption of normality. However, it tends to be restrictive in applications. Data transformation is one alternative way, but it may affect the distinctive characteristics of the original data. Motivated by the data set that exhibits asymmetry, including environmental, financial, and biomedical ones, we propose a high-dimensional discriminant analysis model called the SKNC model (short for SKewed Normal Classification). By incorporating the skewed normal model, the SKNC model inherits all properties of the normal distribution and improves its flexibility on skewed data in classification. Theoretical results rigorously show that the SKNC model achieves variable selection, penalized estimation, and prediction consistency, especially in high-dimensional settings. We empirically demonstrate the superior performance of the SKNC model over existing methods in simulated and real datasets.
