Main Content

ecmmvnrfish

Fisher information matrix for multivariate normal regression model

Syntax

Fisher = ecmmvnrfish(Data,Design,Covariance,Method,MatrixFormat,CovarFormat)

Arguments

Data

NUMSAMPLES-by-NUMSERIESmatrix withNUMSAMPLESsamples of aNUMSERIES-dimensional random vector. Missing values are represented asNaNs. Only samples that are entirelyNaNs are ignored. (To ignore samples with at least oneNaN, usemvnrfish.)

Design

A matrix or a cell array that handles two model structures:

  • IfNUMSERIES = 1,Designis aNUMSAMPLES-by-NUMPARAMSmatrix with known values. This structure is the standard form for regression on a single series.

  • IfNUMSERIES≥1,Designis a cell array. The cell array contains either one orNUMSAMPLEScells. Each cell contains aNUMSERIES-by-NUMPARAMSmatrix of known values.

    IfDesignhas a single cell, it is assumed to have the sameDesignmatrix for each sample. IfDesignhas more than one cell, each cell contains aDesignmatrix for each sample.

Covariance

NUMSERIES-by-NUMSERIESmatrix of estimates for the covariance of the residuals of the regression.

Method

(Optional) Character vector that identifies method of calculation for the information matrix:

  • hessian— Default method. Use the expected Hessian matrix of the observed log-likelihood function. This method is recommended since the resultant standard errors incorporate the increased uncertainties due to missing data.

  • fisher— Use the Fisher information matrix.

MatrixFormat

(Optional) Character vector that identifies parameters to be included in the Fisher information matrix:

  • full— Default format. Compute the full Fisher information matrix for both model and covariance parameter estimates.

  • paramonly— Compute only components of the Fisher information matrix associated with the model parameter estimates.

CovarFormat

(Optional) Character vector that specifies the format for the covariance matrix. The choices are:

  • 'full'— Default method. The covariance matrix is a full matrix.

  • 'diagonal'— The covariance matrix is a diagonal matrix.

Description

Fisher = ecmmvnrfish(Data,Design,Covariance,Method,MatrixFormat,CovarFormat)computes a Fisher information matrix based on current maximum likelihood or least-squares parameter estimates that account for missing data.

Fisheris aNUMPARAMS-by-NUMPARAMSFisher information matrix or Hessian matrix. The size ofNUMPARAMSdepends onMatrixFormatand on current parameter estimates. IfMatrixFormat = 'full',

NUMPARAMS = NUMSERIES * (NUMSERIES + 3)/2

IfMatrixFormat = 'paramonly',

NUMPARAMS = NUMSERIES

Note

ecmmvnrfishoperates slowly if you calculate the full Fisher information matrix.

Examples

SeeMultivariate Normal Regression,Least-Squares Regression,Covariance-Weighted Least Squares,Feasible Generalized Least Squares, and看似不相关的注册ression.

版本历史

Introduced in R2006a

See Also

|

Topics