支持向量分位数变系数模型的估计及应用  

Estimation and application for support vector quantile variable coefficient model

作　　者：聂茹佳 黎文泽 赵为华[1] 张日权 NIE Rujia;LI Wenze;ZHAO Weihua;ZHANG Riquan(School of Sciences,Nantong University,Nantong 226019,China;School of Statistics and Information,Shanghai University of International Business and Economics,Shanghai 201620,China)

机构地区：[1]南通大学理学院,江苏南通226019 [2]上海对外经贸大学统计与信息学院,上海201620

出　　处：《扬州大学学报（自然科学版）》2022年第6期18-25,共8页Journal of Yangzhou University：Natural Science Edition

基　　金：国家自然科学基金资助项目(11971171);国家大学生创新实践资助项目(202210304005Z).

摘　　要：基于不对称“ε-不敏感”损失函数提出变系数模型的参数估计方法,并应用支持向量回归理论和凸二次规划理论给出模型的估计算法.结果显示,所得估计曲线可在不同分位数水平得到一个稀疏解,而数值模拟研究和共享单车实例数据分析结果进一步证实该方法的正确性和实用性.Based on the asymmetric‘ε-insensitive’loss function,the parameter estimation method of the variable coefficient model is proposed,and the estimation algorithm of the model is given using support vector regression theory and convex quadratic programming theory.The results show that a sparse solution of the estimated curve can be obtained at different quantile levels,and the correctness and practicability of the proposed method are further verified by numerical simulation and data analysis of bike-sharing cases.

关 键 词：支持向量回归 分位数回归 变系数模型 “ε-不敏感”损失函数 

分 类 号：O212.2[理学—概率论与数理统计]