Subgroup analysis for functional partial linear regression model

Ma，Haiqiang1; Liu，Chao2; Xu，Sheng3; Yang，Jin4

2022

10.1002/cjs.11696

CANADIAN JOURNAL OF STATISTICS-REVUE CANADIENNE DE STATISTIQUE   影响因子和分区

0319-5724

1708-945X

In a functional partial linear regression (FPLR) model, where the response variable is scalar while the explanatory variables involve both infinite-dimensional functional predictors and finite-dimensional scalar covariates, the relationships between the response and the explanatory variables are often assumed to be the same for all subjects. This article relaxes this assumption and considers a subgroup analysis for the FPLR model, which allows the intercepts to vary for different subgroups from a heterogeneous population. By projecting the functional predictors onto the corresponding eigenspace, the subgroup analysis based on the FPLR model can be simplified to a framework that is similar to the classical subgroup analysis problem. To automatically identify subgroups among observations and estimate the regression parameters of interest, we combine the functional principal component analysis with the concave pairwise penalized approach and develop an ADMM algorithm for functional subgroup analysis. We also establish the consistency of the proposed estimators under mild conditions. Simulation experiments demonstrate that the concave penalized subgroup approach could potentially achieve substantial gains over the ordinary FPLR model. The analysis of data from a creative achievement study is used to illustrate the practical performance of the subgroup analysis for the FPLR model.

Concave penalized approach functional partial linear model functional principal components analysis heterogeneity subgroup analysis

SCI

英语

其他

National Natural Science Foundation of China[11701235,11961028,12161042,11901315] ; Science and Technology Research Project of Education Department of Jiangxi Province["GJJ200522","GJJ200545"] ; China Postdoctoral Science Foundation["2021M691443","2021TQ0141"]

Mathematics

Statistics & Probability

WOS:000769983600001

WILEY

MATHEMATICS

2-s2.0-85126358469

Scopus

1.School of Statistics,Jiangxi University of Finance and Economics,Nanchang,Jiangxi,330013,China
2.Department of Statistics and Data Science,The Southern University of Science and Technology,Shenzhen,Guangdong,518055,China
3.Department of Applied Mathematics,The Hong Kong Polytechnic University,Kowloon,Hong Kong
4.Biostatistics and Bioinformatics Branch,Eunice Kennedy Shriver National Institute of Child Health and Human Development,National Institutes of Health,Bethesda,20817,United States

Ma，Haiqiang,Liu，Chao,Xu，Sheng,et al. Subgroup analysis for functional partial linear regression model[J]. CANADIAN JOURNAL OF STATISTICS-REVUE CANADIENNE DE STATISTIQUE,2022,51(2).

Ma，Haiqiang,Liu，Chao,Xu，Sheng,&Yang，Jin.(2022).Subgroup analysis for functional partial linear regression model.CANADIAN JOURNAL OF STATISTICS-REVUE CANADIENNE DE STATISTIQUE,51(2).

Ma，Haiqiang,et al."Subgroup analysis for functional partial linear regression model".CANADIAN JOURNAL OF STATISTICS-REVUE CANADIENNE DE STATISTIQUE 51.2(2022).