不完全观测的部分函数型线性分位数回归模型及应用  

Partially functional linear quantile regression model and its application for incomplete observations

作　　者：杨玉杰 凌能祥[1] YANG Yujie;LING Nengxiang(School of Mathematics,Hefei University of Technology,Hefei 230009,Anhui,China)

机构地区：[1]合肥工业大学数学学院,安徽合肥230009

出　　处：《山东大学学报(理学版)》2025年第3期100-106,共7页Journal of Shandong University(Natural Science)

基　　金：国家自然科学基金资助项目(72071068)。

摘　　要：首先基于不完全观测的函数型数据,介绍部分函数型线性分位数回归模型及模型的估计方法和实施预测的步骤。其次,由于不完全观测的函数型变量广泛存在,利用尼泊尔2019年4月8日至2020年8月31日期间每10分钟一次的测风塔记录数据进行实证分析。针对风速数据的不完全函数型特征,构建以此为协变量、以日均气压为响应变量的不完全部分函数型线性分位数回归模型,获得模型未知斜率函数和未知参数的估计量,并且对日均气压进行预测分析,进一步说明模型及方法的有效性。Firstly,based on incomplete observed functional data,partially functional linear quantile regression model,the estimation method and prediction step are introduced.Secondly,due to the widespread existence of incomplete observed functional variables,the mast records every 10 minutes from April 8,2019 to August 31,2020 in Nepal for empirical analysis are used.Aiming at the incomplete functional characteristics of wind speed data,an incomplete partial functional linear quantile regression model with daily mean air pressure as the response variable is constructed,and the estimators of the unknown slope function and unknown parameters of the model are obtained,and the daily mean air pressure is predicted and analyzed,which further illustrates the effectiveness of the model and the method.

关 键 词：函数型数据分析 不完全观测 部分函数线性分位数回归 主成分分析 

分 类 号：O212.7[理学—概率论与数理统计]