Mean-Variance Single Change-Point Model

Implementation of the MeanVar-SCP model from Berlind, Cappello, and Madrid Padilla (2025). The function meanvar_scp() takes a length \(T\) vector \(y_{1:T}\) with a single joint mean and variance change and returns the posterior distribution of the change-point.

Usage

meanvar_scp(y, lambda, omega, u_bar, lgamma_u_bar, v, log_pi)

Arguments

y

A numeric vector. \(T\) observations with a single joint mean and variance change.

lambda

A numeric vector. Known trend component of precision of y.

omega

A positive scalar. Prior precision parameter.

u_bar

A numeric vector. Posterior shape parameters equal to \(u_0 + T - t + 1\) for each \(t\).

lgamma_u_bar

A numeric vector. Log gamma function evaluated at u_bar.

v

A numeric vector. Vector of prior rate parameters for each \(t\).

log_pi

A numeric vector. Vector of log prior probabilities for the location of the change-point.

Value

A list. A list of posterior parameters including the mean b_bar, precision omega_bar, rate v_bar, and posterior probabilities of the change-point location pi_bar.