统计与决策2025，Vol.41Issue(24)：64-69,6.DOI:10.13546/j.cnki.tjyjc.2025.24.011
基于自适应Group Lasso的函数型Logistic可加模型的稀疏估计
Sparse Estimation of Functional Logistic Additive Models Based on Adaptive Group Lasso
摘要
Abstract
This paper constructs a functional Logistic additive model for functional covariates and binary response variables,where the influence of functional covariates on the response variable is represented in a non-parametric additive form.Functional principal component analysis(FPCA)is used to reduce dimensionality and compute scaled scores for functional principal compo-nents.The paper proposes an Adaptive Group Lasso method based on B-spline basis functions to simultaneously estimate the co-efficient functions and select nonzero additive terms,and finally obtains the estimation results of relevant variables via smoothing splines.This method is undertaken to mitigate variability,enhance predictive accuracy,and ensure an appropriate level of model fitting.The consistency of the method is strictly proved by establishing theorems,and its consistency and effectiveness are verified through numerical simulation and empirical analysis.关键词
函数型Logistic可加模型/自适应Group Lasso/稀疏估计/平滑样条/相合性Key words
functional Logistic additive models/adaptive Group Lasso/sparse estimation/smooth spline/consistency分类
数理科学引用本文复制引用
Li Chunjing,Wang Yue,Yuan Xiaohui..基于自适应Group Lasso的函数型Logistic可加模型的稀疏估计[J].统计与决策,2025,41(24):64-69,6.基金项目
国家社会科学基金一般项目(24BTJ061) （24BTJ061）
