Composite likelihood estimation in multivariate data analysis

By Yinshan Zhao, Harry Joe. 2005

171 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/44af8ec4e6bc1950a17d4e188689d959fe49cf9f

How researchers see this paper

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

In Statistics and Probability

Citations in the corpus

Copula-Based Models for Multivariate Discrete Response Data (2013, methodology): “Therefore, Genz and Bretz [11] proposed a randomized quasi Monte Carlo method with the use of antithetic variates and Joe [18] proposed two approxima- 1 Both approximations to MVN rectangle in [18], the 1-dimensional integral in the exchangeable case, and the method in [50], can be computed with the”
ISSUES AND STRATEGIES IN THE SELECTION OF COMPOSITE LIKELIHOODS (2011, background): “…likelihood, has received increasing attention in recent years due to the simplicity in defining the objective function and computational advantages when dealing with data with complex structure (e.g., Kuk and Nott (2000); Zhao and Joe (2005); Fieuws and Verbeke (2006); Dillon and Lebanon (2009)).”
Limited Information Goodness-of-fit Testing in Multidimensional Contingency Tables (2006, unknown): “Also, we use a bivariate composite likelihood (BCL) estimator (Zhao & Joe, 2005) under the same conditions as above to illustrate the behavior of M2 for estimators that are not BAN.”
A New Generalized Heterogeneous Data Model (GHDM) to Jointly Model Mixed Types of Dependent Variables (2015, methodology): “(log)( 1 , Q q qMACMLMACML LL δδ (22) The covariance matrix of the parameters δ may be estimated by the inverse of Godambe’s (1960) sandwich information matrix (see Zhao and Joe, 2005; Bhat, 2014).”
AN OVERVIEW OF COMPOSITE LIKELIHOOD METHODS (2011, methodology): “The weights 1/(mi−1), recommended in Kuk and Nott (2000), LeCessie and van Houwelingen (1994), and Zhao and Joe (2005), are suitable for the limiting case of independence within clusters, but the weights 1/{mi(mi−1)} are optimal for very strong dependence.”
A comparison of the maximum simulated likelihood and composite marginal likelihood estimation approaches in the context of the multivariate ordered-response model (2010, methodology): “…efficiency from a theoretical perspective relative to a full likelihood estimator, but this efficiency loss appears to be empirically minimal (see Zhao and Joe, 2005; Lele, 2006; Joe and Lee, 2009).2 Besides, the simulation estimation methods for evaluating the analytically intractable…”
On composite likelihood estimation of a multivariate INAR(1) model (2013, methodology): “Thus, the window subsampling method of Heageart and Lele (1998), Monte Carlo integration or resampling methods such as jackknife or bootstrap can also be considered for a more robust evaluation of the standard errors (see e.g. Lipsitz et al., 1994; Zhao and Joe, 2005).”
Pair Copula Constructions for Multivariate Discrete Data (2012, background): “Zhao and Joe ( 2005 ) provide a faster alternative to estimate the parameters based on composite likelihood methods, however, this comes at the cost of statistical efficiency when compared to maximum likelihood estimation.”
The Composite Marginal Likelihood (CML) Inference Approach with Applications to Discrete and Mixed Dependent Variable Models (2014, unknown): “An alternative procedure to obtain the covariance matrix of the CML estimator is to use a jackknife approach as follows (see Zhao and Joe, 2005): ( )( ) ,θθθθθ CMLCMLCMLCMLCML ′−−−= −− = ∑ ˆˆˆˆ1)ˆ(Cov )()( 1 qq Q qQ Q (1.18) where )(ˆ q−CMLθ is the CML estimator with the qth observational unit…”
A New Estimation Approach to Integrate Latent Psychological Constructs in Choice Modeling (2014, methodology): “(log)( 1 , = = Q q qMACMLMACML LL θθ 5 The covariance matrix of the parameters θ may be estimated by the inverse of Godambe’s (1960) sandwich information matrix (see Zhao and Joe, 2005).”
A latent variable representation of count data models to accommodate spatial and temporal dependence: application to predicting crash frequency at intersections (2011, background): “…asymptotically normal distributed with asymptotic mean θ and covariance matrix given by the inverse of Godambe’s (1960) sandwich information matrix (see Zhao and Joe, 2005): 111 )]()[()]([)]([)ˆ( −−− == θHθJθHθGθVCML , (6) where ⎥ ⎦ ⎤ ⎢ ⎣ ⎡ ′∂∂ ∂ −= θθ θθH )(log)( 2 CMLLE and ⎥ ⎦ ⎤ ⎢ ⎣ ⎡ ⎟ ⎠ ⎞ ⎜…”
The maximum approximate composite marginal likelihood (MACML) estimation of multinomial probit-based unordered response choice models (2011, methodology): “…loss of the CML estimator (relative to the maximum likelihood (ML) estimator) appears to be negligible to small from an empirical standpoint (see Zhao and Joe, 2005; Lele, 2006; Joe and Lee, 2009).7 In the next few sections, we discuss how the MVNCD function approximation of Section 2.1 and the…”
A robust pairwise likelihood method for incomplete longitudinal binary data arising in clusters (2011, methodology): “…of using quasi likelihoods based on PWL has received increasing attention in recent years due to the simplicity in defining the objective function and computational advantages when dealing with data with complex structure (e.g., Heagerty & Lele, 1998; Zhao & Joe, 2005; Fieuws & Verbeke, 2006).”
Gaussian Copula Marginal Regression (2012, methodology): “Whenever this integral factorizes in low-dimensional terms as in longitudinal studies with few observations per subject, we suggest to use precise deterministic approximations as the method of Joe [27] or other recent numerical methods by Miwa et al. [38] and Craig [9].”
Joint composite estimating functions in spatiotemporal models (2012, methodology): “Zhao and Joe (2005) used a similar approach to account for different cluster sizes in their CL formulation for familial data.”