deriv_prior: Gradient for priors

In thiyangt/fformpp: Feature-based FORecast Model Performance Prediction

Gradient for priors

Description

A collection of gradient for common priors.

Usage

deriv_prior(B, priorArgs, hessMethod)

Arguments

B	"matrix". The paramter that need to be added with a prior. The gradient and hessian are calculated conditional on B. B should be always an one-column matrix,
priorArgs	"list". priorArgs$prior_type: when prior_type is set to "mvnorm", you have to provide priorArgs$mean: "matrix", the mean of parameter, mu0 should be always an one-column matrix; priorArgs$covariance: "matrix", the covariance matrix. A g-prior can be constructed by setting it to X'X, where X is the covariates matrix.; priorArgs$shrinkage: "numeric", the shrinkage for the covariance.

Details

The parameters after "..." should be matched exactly.

Value

"list". The gradient and hessian matrix, see below.

Note

First version: Tue Mar 30 09:33:23 CEST 2010; Current: Wed Sep 15 14:39:01 CEST 2010. TODO:

Author(s)

Feng Li, Department of Statistics, Stockholm University, Sweden.

thiyangt/fformpp documentation built on Jan. 5, 2024, 5:44 a.m.