bsm.fit {msos}R Documentation

Helper function to determine β\beta estimates for MLE regression with patterning.

Description

Generates β\beta estimates for MLE using a conditioning approach with patterning support.

Usage

bsm.fit(x, y, z, pattern)

Arguments

x

An N×(P+F)N \times (P + F) design matrix, where FF is the number of columns conditioned on. This is equivalent to the multiplication of xyzbxyzb.

y

The N×(QF)N \times (Q - F) matrix of observations, where FF is the number of columns conditioned on. This is equivalent to the multiplication of YzaYz_a.

z

A (QF)×L(Q - F) \times L design matrix, where FF is the number of columns conditioned on.

pattern

An optional NFxFN-F x F matrix of 0's and 1's indicating which elements of β\beta are allowed to be nonzero.

Value

A list with the following components:

Beta

The least-squares estimate of β\beta.

SE

The (P+F)×L(P+F)\times L matrix with the ijijth element being the standard error of β^ij\hat{\beta}_ij.

T

The (P+F)×L(P+F)\times L matrix with the ijijth element being the t-statistic based on β^ij\hat{\beta}_ij.

Covbeta

The estimated covariance matrix of the β^ij\hat{\beta}_ij's.

df

A pp-dimensional vector of the degrees of freedom for the tt-statistics, where the jjth component contains the degrees of freedom for the jjth column of β^\hat{\beta}.

Sigmaz

The (QF)×(QF)(Q - F) \times (Q - F) matrix Σ^z\hat{\Sigma}_z.

Cx

The Q×QQ \times Q residual sum of squares and crossproducts matrix.

See Also

bothsidesmodel.mle and bsm.simple

Examples

# NA

[Package msos version 1.2.0 Index]