混合偏正态数据下中位数回归模型的参数估计

Estimation for Skew-Normal Mixture of Median Regression Models

有限混合回归模型是处理来自异质总体数据的重要工具.偏斜/不对称数据经常出现在金融、经济、社会科学、气候科学、环境科学、工程、生物医学等领域.根据偏正态数据的特征,均值易受极端值的影响,而中位数能代表“中等”水平,在偏态数据下比均值更具代表性.作者介绍了基于混合偏正态数据下的中位数回归模型,通过一种改进的EM算法进行参数估计,模拟研究和实例分析表明本文提出的方法是有效的.

Finite mixture of regression models are one of the most important statistical data analysis tools to handle heterogeneous populations.Skewed/asymmetric data frequently appear in research fields such as finance,econom-mics,social sciences,climate science,environmental science,engineering,and biomedicine.According to the characteristics of skew-normal data,the mean is susceptible to the influence of extreme values,while the median represents the“medium”level,which is more representative than the mean under skewed data.In this paper,the authors introduce the median regression model based on mixture skew-normal data,and use a modified EM algorithm to estim-ate the parameters.The methodology is illustrated through numerical experiments and a real data example.