报告主题：A Penalized Likelihood Method for Switching Regressions with Skew-normal Errors

时间：2022年10月21日13:00-13:45

地点：腾讯会议-会议号：369 2414 5327

报告人：金立斌

报告内容简介：

This paper establishes identifiability results for the switching regression models with skew normal errors under fixed and random designs. In addition, we propose a novel penalized maximum likelihood estimation method in the model and show the strong consistency of the proposed estimator. An EM-type algorithm for deriving the penalized estimator is presented. The finite sample properties of the proposed methodology are studied through simulations and a real data example is used for illustration.