Problem
Research questions and friction points this paper is trying to address.
multi-source data integration
joint-individual separation
joint rank estimation
scalability
Testing and segmentation of joint and individual components in integrative multi-source factor models
🤖 AI Summary
This study addresses the challenges of low separation accuracy and high computational complexity in disentangling shared and source-specific structures during multi-source data integration. To this end, we propose MSSAT, a method grounded in the geometric insights of subspace alignment that integrates multi-source factor models with spectral analysis. By deriving the asymptotic null distribution, MSSAT establishes a resampling-free sequential hypothesis testing framework for efficient joint rank estimation. Extensive evaluations on both simulated datasets and TCGA multi-omics data demonstrate that MSSAT significantly improves structural disentanglement accuracy while substantially reducing computational overhead. Ultimately, this work provides a rigorous and computationally efficient statistical inference tool for high-dimensional multi-source data fusion.
📝 Abstract
Disentangling shared (joint) structures from source-specific (individual) variations is a fundamental task in multi-source data integration. Existing joint-individual models often rely on computationally intensive optimization or loose spectral bounds, leading to suboptimal separation accuracy and poor scalability. In this paper, we propose the Multi-Source Sequential Alignment Test (MSSAT). MSSAT leverages the geometric observation that true joint components manifest as closely aligned score subspaces across different data sources. By deriving the asymptotic null distribution of our alignment statistic, we develop a rigorous, resampling-free sequential testing procedure to accurately estimate the joint rank. Extensive simulations and real data applications, including a TCGA multi-omics dataset, demonstrate that MSSAT achieves superior separation accuracy and substantially faster computation compared to competing methods.
Innovation
Methods, ideas, or system contributions that make the work stand out.
multi-source data integration
joint-individual decomposition
sequential alignment test
subspace alignment
asymptotic null distribution
🔎 Similar Papers
No similar papers found.
