Researchers have developed a new statistical method for testing conditional independence in the graphical structure of stationary Gaussian time series, particularly addressing the challenge of long-memory data. This approach utilizes a data-adaptive test statistic and establishes a finite-sample Gaussian approximation bound applicable to both short- and long-memory series. The procedure is robust, employing a block bootstrap method for validity even in ultra-high-dimensional scenarios, and includes a consistency-empowering correction to ensure asymptotic accuracy in size and power. The method was successfully applied to real-world functional magnetic resonance imaging (fMRI) data to analyze functional connectivities within the brain. AI
IMPACT This research advances statistical methods for analyzing complex time series data, potentially improving the accuracy of models used in fields like neuroscience.
RANK_REASON The cluster contains a research paper detailing a new statistical methodology. [lever_c_demoted from research: ic=1 ai=0.4]
- Berry--Esseen Bounds and ASCLTs for Drift Parameter Estimator of Mixed Fractional Ornstein--Uhlenbeck Process with Discrete Observations
- Block-bootstrapping for noisy data
- functional magnetic resonance imaging
- Gaussian Graphical Model
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
