eba: Extreme bounds analysis summaries from a Bayesian model space

In rmsBMA: Reduced Model Space Bayesian Model Averaging

Extreme bounds analysis summaries from a Bayesian model space

Description

Computes Extreme Bounds Analysis (EBA) summaries for the intercept and each regressor across a model space. For each coefficient, the function reports: the minimum coefficient ("Low"), maximum coefficient ("High"), the mean coefficient ("Mean_coef"), and corresponding "extreme bounds" defined as \mathrm{Low} - 2\cdot \mathrm{SE} and \mathrm{High} + 2\cdot \mathrm{SE}, where \mathrm{SE}=\sqrt{\mathrm{VAR}} is the standard error associated with the coefficient estimate in the model attaining the minimum/maximum.

Usage

eba(betas, VAR, Reg_ID, var_tol = 0)

Arguments

betas	Numeric matrix of dimension MS x (K+1) containing estimated coefficients across models. Column 1 corresponds to the intercept, columns 2 to K+1 correspond to regressors.
VAR	Numeric matrix of dimension MS x (K+1) containing variances of the coefficient estimates. Must have the same dimensions as betas.
Reg_ID	Numeric or integer matrix of dimension MS x K indicating regressor inclusion. Entry Reg_ID[i,k]=1 if regressor k is included in model i, and 0 otherwise.
var_tol	Nonnegative numeric scalar used as a tolerance when checking variance positivity. Entries with VAR <= var_tol are treated as invalid for bound calculations. Default is 0.

Details

The intercept (constant) is assumed to be included in all models. Each regressor is summarized only over models in which it is included, as indicated by the model-inclusion matrix Reg_ID.

Value

A numeric matrix of dimension (K+1) x 5 with columns:

Lower_bound

\min(\beta) - 2\cdot \mathrm{SE} evaluated at the model where \beta is minimal.

Low

Minimum coefficient value across relevant models.

Mean_coef

Mean coefficient across relevant models (intercept: all models; regressor: included models only).

High

Maximum coefficient value across relevant models.

Upper_bound

\max(\beta) + 2\cdot \mathrm{SE} evaluated at the model where \beta is maximal.

Rows correspond to the intercept (row 1) and regressors (rows 2..K+1). If a regressor is never included (no 1s in its column of Reg_ID), its row will contain NA.

rmsBMA documentation built on March 14, 2026, 5:06 p.m.