基于复合分位数回归的变系数模型平均  

Average of Variable Coefficient Models Based on Compound Quantile Regression

作　　者：谭蓉 TAN Rong(School of Mathematical Sciences,Chongqing Normal University,Chongqing 401331,China)

机构地区：[1]重庆师范大学数学科学学院,重庆401331

出　　处：《兰州文理学院学报(自然科学版)》2025年第1期29-34,共6页Journal of Lanzhou University of Arts and Science(Natural Sciences)

基　　金：国家自然科学基金项目(12201091);重庆市自然科学基金面上项目(CSTB2022NSCQ-MSX0852);全国统计科学研究项目(2022LY019);重庆市教育委员会科学技术研究项目(KJQN202100526)。

摘　　要：针对变系数模型的平均问题,采用一种新的半参数建模策略,即通过轮换连续变量作为指标变量的方式构建一系列非嵌套的候选模型.基于复合分位数回归估计候选模型,运用弃一交叉验证准则选择权重进行模型平均预测.模拟研究表明该方法具有更好的有限样本性质,最后将该方法应用到Boston住房数据,进一步说明其准确性和有效性.To address the problem of averaging variable coefficient models,a new semi-parametric modeling strategy is used,i.e.,a series of non-nested candidate models are constructed by rotating continuous variables as indicator variables.The candidate models are estimated based on composite quantile regression,and the discard-one cross-validation criterion is applied to select weights for model averaging prediction.A simulation study demonstrates the better finite-sample nature of the method,and finally the method is applied to Boston housing data to further illustrate its accuracy and validity.

关 键 词：复合分位数回归 模型平均 变系数模型 

分 类 号：O212.1[理学—概率论与数理统计]