Robust EM Type Algorithm for Mixture Regression Models with Censored Data

• 作 者: Yaqin Xing Xiuqing Zhou
• 机 构: School of Mathematical Sciences and Institute of Finance and Statistics,Nanjing Normal University,Na
• 出 处: 中国现场统计研究会
• 发表日期: 2015年3期
• 关键词: Censored data Mixture regression models EM algorithm M-estimation

　　The maximum likelihood estimation(MLE)and the least squares estimation have been widely used for fitting the mixture regression models,and robust methods have also been developed to avoid the sensitivity of the estimation methods in the case where exist outliers or the error distribution has heavy tails.