A New Parametrization of Correlation Matrices

@article{Archakov2020ANP,
  title={A New Parametrization of Correlation Matrices},
  author={Ilya Archakov and Peter Reinhard Hansen},
  journal={arXiv: Econometrics},
  year={2020}
}
We introduce a novel parametrization of the correlation matrix. The reparametrization facilitates modeling of correlation and covariance matrices by an unrestricted vector, where positive definiteness is an innate property. This parametrization can be viewed as a generalization of Fisther's Z-transformation to higher dimensions and has a wide range of potential applications. An algorithm for reconstructing the unique n x n correlation matrix from any d-dimensional vector (with d = n(n-1)/2) is… 

Figures and Tables from this paper

A Canonical Representation of Block Matrices with Applications to Covariance and Correlation Matrices.
We obtain a canonical representation for block matrices. The representation facilitates simple computation of the determinant, the matrix inverse, and other powers of a block matrix, as well as the
Theoretically and computationally convenient geometries on full-rank correlation matrices
In contrast to SPD matrices, few tools exist to perform Riemannian statistics on the open elliptope of full-rank correlation matrices. The quotient-affine metric was recently built as the quotient of
A Multivariate Realized GARCH Model
TLDR
A factor approach offers a way to impose a parsimonious structure in high dimensional system and it is shown that a factor framework arises naturally in some existing models.
Regression Analysis of Correlations for Correlated Data
TLDR
A novel regression approach for modelling the correlation structure is proposed, leveraging a new generalized z-transformation, which maps correlation matrices that are constrained to be positive definite to vectors with un-restricted support, and is order-invariant.
Multivariate Stochastic Volatility Models based on Generalized Fisher Transformation*
Modeling multivariate stochastic volatility (MSV) faces a few challenges when both variances and covariances are time-varying. The main challenges come from the assurance of positive-definiteness of
Gaussian Pixie Autoencoder: Introducing Functional Distributional Semantics to continuous latent spaces
TLDR
A new Gaussian Pixie Autoencoder model is proposed to introduce FDS to continuous latent modelling, improves on the predecessors in terms of simplicity and efficiency setting a new baseline for continuous FDS models.
Dynamic factor, leverage and realized covariances in multivariate stochastic volatility
In the stochastic volatility models for multivariate daily stock returns, it has been found that the estimates of parameters become unstable as the dimension of returns increases. To solve this
Macroeconomic Forecasting and Variable Ordering in Multivariate Stochastic Volatility Models
We document five novel empirical findings on the well-known potential ordering drawback associated with the time-varying parameter vector autoregression with stochastic volatility developed by Cogley
Semiparametric Gaussian Copula Regression modeling for Mixed Data Types (SGCRM)
TLDR
Semiparametric Gaussian Copula Regression modeling (SGCRM) is proposed that allows to model a joint dependence structure between observed continuous, truncated, ordinal, and binary variables and to construct conditional models with these four data types as outcomes with a guarantee that derived conditional models are mutually consistent.
...
...

References

SHOWING 1-10 OF 42 REFERENCES
The Matrix-Logarithmic Covariance Model
Abstract A flexible method is introduced to model the structure of a covariance matrix C and study the dependence of the covariances on explanatory variables by observing that for any real symmetric
Unconstrained parametrizations for variance-covariance matrices
TLDR
Five different parametrizations for variance-covariance matrices that ensure positive definiteness, thus leaving the estimation problem unconstrained in maximum likelihood and restricted maximum likelihood estimation in linear and non-linear mixed-effects models are described.
Covariance Estimation: The GLM and Regularization Perspectives
TLDR
A survey of the progress made in modeling covariance matrices from two relatively complementary perspectives: (1) generalized linear models (GLM) or parsimony and use of covariates in low dimensions, and (2) regularization or sparsity for high-dimensional data.
A Multivariate Realized GARCH Model
TLDR
A factor approach offers a way to impose a parsimonious structure in high dimensional system and it is shown that a factor framework arises naturally in some existing models.
Penalized Covariance Matrix Estimation Using a Matrix-Logarithm Transformation
TLDR
A generalization of the Leonard and Hsu log-likelihood approximation that no longer requires a nonsingular sample covariance matrix is developed and can be solved efficiently by an iterative quadratic programming algorithm.
Matrix exponential GARCH
Interpreting the First Eigenvalue of a Correlation Matrix
The first eigenvalue of a correlation matrix indicates the maximum amount of the variance of the variables which can be accounted for with a linear model by a single underlying factor. When all
Dynamic conditional correlation models for realized covariance matrices
New dynamic models for realized covariance matrices are proposed. The expected value of the realized covariance matrix is specified in two steps: one for each realized variance, and one for the
...
...