RT Generic
T1 D-Trace precision matrix estimator with eigenvalue control
A1 Avagyan, Vahe
A2 Universidad Carlos III de Madrid. Departamento de Estadística, 
AB The estimation of a precision matrix has an important role in several research fields. In high-dimensional settings, one of the most prominent approaches to estimate theprecision matrix is the ɭ₁ (Lasso) norm penalized convex optimization. This frameworkguarantees the sparsity of the estimated precision matrix. However, it does not controlthe eigenspectrum of the obtained estimator, and, moreover, it shrinks the largesteigenvalues of the estimated precision matrix. In this paper, we focus on D-traceprecision matrix methodology. We propose imposing a negative trace penalization onthe objective function of the D-trace approach, aimed to control the eigenvalues.Through extensive numerical analysis, using simulated and real datasets, we show theadvantageous performance of our proposed methodology.
SN 2387-0303
YR 2016
FD 2016-07
LK https://hdl.handle.net/10016/23410
UL https://hdl.handle.net/10016/23410
LA eng
DS e-Archivo
RD 19 may. 2024