A spectral decomposition approach for multi-view covariance estimation yields conjugate normal-inverse-gamma posteriors and high-dimensional asymptotic guarantees.
Recall that, with probability at least 1 โ ๐(1), || ๐น0๐|| โ โ๐ by Corollary 5.35 of Vershynin (2012) and max๐=1,..., ๐ ๐ || ๐0๐ ๐ || โค || ฮ0||โ โ๐ ๐ โ 1 by Assumption
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ME 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Inference on covariance structure in high-dimensional multi-view data
A spectral decomposition approach for multi-view covariance estimation yields conjugate normal-inverse-gamma posteriors and high-dimensional asymptotic guarantees.