A0179
Title: Robust regression with a nonparametric maximum likelihood approach for skew-normal scale mixtures
Authors: Sangkon Oh - Pukyong National University (Korea, South) [presenting]
Byungtae Seo - Sungkyunkwan University (Korea, South)
Abstract: The ordinary least squares method is commonly used for estimating regression coefficients in linear regression. However, this approach is highly sensitive to influential outliers and loses efficiency when the error follows a skewed or heavy-tailed distribution. To address these limitations, the adoption of semiparametric skew-normal scale mixture distributions for the error is proposed. By using a nonparametric maximum likelihood estimator for the scale factor in the skew-normal scale mixture, the risk of misspecification issues is mitigated, and robust estimators are produced without any model selection procedure. Furthermore, simulation studies and real data analyses are presented to demonstrate the performance of the proposed method.
