Parsimonious Covariance Matrix Estimation for Longitudinal Data

By M. S. Smith, Robert Kohn. 2002

207 citations recorded by the source corpus.

Subfields: Economics and Econometrics, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/784ea6b5d55d28585e8f6cf4b552abcb1edfd995

How researchers see this paper

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

In Economics and Econometrics

In Statistics and Probability

Citations in the corpus

Bayesian Modeling and Forecasting of Intraday Electricity Load (2003, methodology): “Previous analyses (Smith and Kohn 2002) have suggested that this intraday residual correlation is likely to be well modeled by a short-lag autoregressive-stylemodel with autoregressivecoef cients that differ over half hours.”
High dimensional covariance matrix estimation using a factor model (2007, unknown): “See, for example, Dempster (1972), Leonard and Hsu (1992), Chiu, Leonard and Tsui (1996), Diggle and Verbyla (1998), Pourahmadi (2000), Boik (2002), Smith and Kohn (2002), Wong, Carter and Kohn (2003), Wu and Pourahmadi (2003), Huang, Liu and Pourahmadi (2004), and Li and Gui (2005).”
Covariance matrix selection and estimation via penalised normal likelihood (2006, methodology): “Our approach is also related to a Bayesian approach proposed by Smith & Kohn (2002) which places a hierarchical prior to allow zero entries in T . Ledoit & Wolf (2004) considered shrinkage estimation of covariance matrices in a way rather different from our approach.”
Implications of oil prices shocks for the major emerging economies: A comparative analysis of BRICS (2018, unknown): “covariance matrix as presented in equation-4 is appropriate particularly when one is focused on efficiently estimating covariance matrices (see Smith and Kohn 2002, Primiceri 2005). A fairly similar decomposition was carried out by Cogley (2003) and Cogley and Sargent (2003) employing a Time-Varying”
Sparsistency and Rates of Convergence in Large Covariance Matrix Estimation (2007, methodology): “d, whereas in Gaussian graphical models, the sparsity of the precision is a reasonable assumption ([5]). This initiates a series of research focusing on the parsimony estimation of covariance matrix. [18] used prior which admit zeros on the oﬀ-diagonal elements of the Cholesky factor of the precisio”
Bayesian semi parametric multi-state models (2008, background): “These problems can be solved in normal models by marginalisation (compare for example, Smith and Kohn (2002) or Früehwirth-Schnatter & Tüchler (forthcoming)) but are more difficult to handle in non-Gaussian models.”
Sparse Variational Analysis of Linear Mixed Models for Large Data Sets. (2011, unknown): “For the standard model, the priors are specified as α ~ N (0, A0), σ−2 ~ G(c0, d0) and D ~ IW(n0,Ψ0) where α0 = 0, A0 = 1000I (Chen and Dunson, 2003), c0 = 0.1, d0 = 0.001 (Smith and Kohn, 2002), n0 = q and Φ0 = I to reflect our vague prior information on α, σ−2 and D.”
An Hourly Periodic State Space Model for Modelling French National Electricity Load (2008, background): “In our paper we follow Smith & Kohn (2002) in allowing cross-correlation between the stochastic terms of the equations for the different hours of the day.”
Penalising Model Component Complexity: A Principled, Practical Approach to Constructing Priors (2014, methodology): “; Huang and Wand, 2013). An alternate method is to note that every covariance matrix can be decomposed as = LLT, where Lis a lower triangular matrix, and specify a prior directly on the elements of L(Smith and Kohn, 2002; Fruhwirth-¨ Schnatter and Tuchler, 2008). This formulation has the distinct di”
Bayesian stochastic search for VAR model restrictions (2008, methodology): “In particular, Smith and Kohn (2002), who developed an alternative approach to finding parsimony in R À1 , used the likelihood function to the power of the reciprocal of the sample size as the prior on the elements of the Cholesky decomposition of the covariance matrix.”
Adaptive Shrinkage in Bayesian Vector Autoregressive Models (2019, methodology): “Several alternative factorizations of the variance-covariance matrix are available (see, among others, Pourahmadi 1999; Smith and Kohn 2002).”
Stochastic model specification search for Gaussian and partial non-Gaussian state space models (2010, methodology): “It allows parsimonious covariance modelling for longitudinal data as shown by Smith and Kohn (2002) and covariance selection in random effects models as shown by Chen and Dunson (2003) and Frühwirth-Schnatter and Tüchler (2008).”
A review of Bayesian variable selection methods: what, how and which (2009, methodology): “common over all js) (Iswaran and Rao 2005) or analytically integrate it out of the analysis (Smith and Kohn 2002).”