Intrinsic partial linear models for manifold-valued data

作者：

Highlights：

• An intrinsic partial linear modelling (IPLM) framework on the Riemannian manifold is proposed.

• An alternatively iterative strategy is proposed in IPLM.

• The experimental results suggest IPLM leads to better adaptability, interpretation, fitting, and robustness.

摘要

•An intrinsic partial linear modelling (IPLM) framework on the Riemannian manifold is proposed.•An alternatively iterative strategy is proposed in IPLM.•The experimental results suggest IPLM leads to better adaptability, interpretation, fitting, and robustness.

论文关键词：primary,62F12,secondary,58A99,Functional data analysis,Partial linear model,Riemannian manifold,Semi-parametric regression

论文评审过程：Received 3 November 2021, Revised 2 April 2022, Accepted 21 April 2022, Available online 16 May 2022, Version of Record 16 May 2022.

论文官网地址：https://doi.org/10.1016/j.ipm.2022.102954