A dimension-agnostic neural network jointly learns lag transforms and eigenvalue regularization to produce minimum-variance equity portfolios that outperform non-linear shrinkage estimators in 2000-2024 out-of-sample tests.
Nonparametric eigenvalue-regularized precision or covariance matrix estimator.Annals of Statistics, 44(3):928–953
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
q-fin.PM 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
End-to-End Large Portfolio Optimization for Variance Minimization with Neural Networks through Covariance Cleaning
A dimension-agnostic neural network jointly learns lag transforms and eigenvalue regularization to produce minimum-variance equity portfolios that outperform non-linear shrinkage estimators in 2000-2024 out-of-sample tests.