國立中正大學數學系

暨應用數學碩士班、統計科學碩士班

學    術   演    講

A harmonic Mean Loss Function For

Dichotomous/ Polytomous Regression Models and

 Their Misspecification

林 賢一 教授

Prof. Kenichi Hayashi

慶應義塾大学 理工学部 数理科学科

Department of Mathematics, Keio University

Abstract

        The regression models for the categorical response are frequently used in medical research.

      These models assume that the data are sampled from a specific population, but this is often violated in practice. This study introduces a novel loss function that are based on a divergence we call the harmonic mean divergence. While there are many studies on robust methods for regression models, divergence-based approaches have great promise. We show interesting properties of our proposed loss function and its robustness.

日  期:113年10月9日(星期三) 16:10~17:00

地  點：本校數學館527教室（嘉義縣民雄鄉大學路168號）