hessian: Compute Hessian matrix at the MLE

In corncob: Count Regression for Correlated Observations with the Beta-Binomial

Compute Hessian matrix at the MLE

Description

Compute Hessian matrix at the MLE

Usage

hessian(mod, numerical = FALSE)

Arguments

mod	an object of class bbdml
numerical	Boolean. Defaults to FALSE. Indicator of whether to use the numeric Hessian (not recommended).

Value

Hessian matrix at the MLE. In this setting, it's hard to compute expectations to calculate the information matrix, so we return the consistent estimate using sample moments: \hat{A}(\hat{\theta}) = \sum_i \frac{\partial^2}{\partial \theta \partial \theta^T} l(\theta, W_i) evaluated at \theta = \hat{\theta}.

Examples

data(soil_phylum_small_otu1)
mod <- bbdml(formula = cbind(W, M - W) ~ DayAmdmt,
phi.formula = ~ DayAmdmt,
data = soil_phylum_small_otu1)
hessian(mod)

corncob documentation built on May 29, 2024, 7:15 a.m.