HIGHER-ORDER ACCURATE, POSITIVE SEMIDEFINITE ESTIMATION OF LARGE-SAMPLE COVARIANCE AND SPECTRAL DENSITY MATRICES

By Dimitris Nicolas Politis. Econometric Theory, 2005

127 citations recorded by the source corpus.

Subfields: Applied Mathematics, Computer Vision and Pattern Recognition, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/7f234f22b6434f327d7199c6d844329791d40782

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Applied Mathematics

In Computer Vision and Pattern Recognition

In Statistics and Probability

Citations in the corpus

Banded and tapered estimates for autocovariance matrices and the linear process bootstrap (2010, background): “The LPB is also more general because one could, in principle, use a taper j(Æ) that is not identically zero after a point, but just tends to zero, [see, for example Politis (2007)]; in that case the LPB generates linear MA(1) rather than MA(q) processes.”
Dynamic functional principal components (2012, methodology): “For more details concerning common choices of w and the tuning parameter qn, we refer to chapters 10 and 11 in Brockwell and Davis (1981) and to Politis (2011).”
Testing stationarity of functional time series (2014, unknown): “Politis (2003, 2011) argues that the flat top kernel K (t) =  1, 0 ≤ t 0.1 1.1 − |t|, 0.1 ≤ t 1.1 0, |t| ≥ 1.1 (4.1) has better properties than the Bartlett or the Parzen kernels.”
Covariance matrix estimation and linear process bootstrap for multivariate time series of possibly increasing dimension (2015, methodology): “They make use of an approach primarily proposed in Politis (2003b) to estimate the bandwidth in spectral density estimation which has been generalized to the multivariate case in Politis (2011).”
Change-point detection in panel data via double CUSUM statistic (2016, methodology): “We estimate σj using the flat-top kernel estimator with the automatically chosen bandwidth as discussed in Politis (2011). Let η̂j,1 = arg maxb∈[1,T ) |Cb({xj,t}t=1)|, ε̄j,t = xj,t − 1 η̂j,1 η̂j,1 ∑”
TFT-bootstrap: Resampling time series in the frequency domain to obtain replicates in the time domain (2011, methodology): “Bühlmann [6], Lahiri [36] and Politis [50] give reviews of the state-of-the-art in resampling time series and dependent data.”