Spatial Kernel Restricted Canonical Correlation Analysis: Population Targets, Identifiability, and Stability
Abstract
Relating observations of different categories across space requires methods that accommodate distinct sampling locations and spatial dependence. Spatial kernel restricted canonical correlation analysis (skrCCA) uses spatial weights to identify associated feature combinations across data categories. Here, we develop a statistical theory of its covariance form after feature reduction. Conditional on the observed locations and under a mean-zero model, the pairwise association matrices target kernel-weighted averages of cross-covariances. For prespecified directions, we derive the exact null variance under independence between blocks. We establish recovery and stability bounds in two- and multiblock settings under explicit population models, and show that exact cyclic updates have convergent objective values. The variance characterization shows how spatial weighting and within-block dependence jointly determine the effective information available for estimating association. The recovery bounds separate statistical and optimization error, relating stability to population separation and the strength of connections between blocks. These results clarify what fitted spatial association axes estimate and provide conditions under which they can be recovered reliably.
Related
- Canonical correlation analysis identifies single nucleotide polymorphism combinations and gene expression associated with membranous nephropathy
- Belowground functional traits predict population temporal stability in grasslands
- Revisiting the evolution of population stability due to selection for rapid development and early reproduction in Drosophila: the role of generation length
- Identifiability of metabolic resilience from sparse longitudinal metabolomics
- Identifiability of a Simple Model of Lateral Gene Transfer
BibTeX
@misc{miao2026spatial,
title = {{Spatial Kernel Restricted Canonical Correlation Analysis: Population Targets, Identifiability, and Stability}},
author = {Zhen Miao and Yilong Qu and Nancy R. Zhang},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.17830},
eprint = {2610.17830},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.17830},
}
RIS
TY - UNPB
TI - Spatial Kernel Restricted Canonical Correlation Analysis: Population Targets, Identifiability, and Stability
AU - Zhen Miao
AU - Yilong Qu
AU - Nancy R. Zhang
PY - 2026
DA - 2026/10/13
PB - arXiv
DO - 10.48550/arxiv.2610.17830
UR - https://arxiv.org/abs/2610.17830
AB - Relating observations of different categories across space requires methods that accommodate distinct sampling locations and spatial dependence. Spatial kernel restricted canonical correlation analysis (skrCCA) uses spatial weights to identify associated feature combinations across data categories. Here, we develop a statistical theory of its covariance form after feature reduction. Conditional on the observed locations and under a mean-zero model, the pairwise association matrices target kernel-weighted averages of cross-covariances. For prespecified directions, we derive the exact null variance under independence between blocks. We establish recovery and stability bounds in two- and multiblock settings under explicit population models, and show that exact cyclic updates have convergent objective values. The variance characterization shows how spatial weighting and within-block dependence jointly determine the effective information available for estimating association. The recovery bounds separate statistical and optimization error, relating stability to population separation and the strength of connections between blocks. These results clarify what fitted spatial association axes estimate and provide conditions under which they can be recovered reliably.
ER -