peikalunci/dynamic-probit-ep

Expectation approximation for the smoothing distribution in dynamic probit models

Expectation propagation for the smoothing distribution in dynamic probit

This repository is associated with the article Anceschi, Fasano and Rebaudo (2023) Expectation propagation for the smoothing distribution in dynamic probit. Bayesian Statistics, New Generations New Approaches (BaYSM2022). The key contribution of the paper is outlined below.

we develop an expectation propagation (EP) approximation of the joint smoothing distribution in a dynamic probit model.

This repository provides codes to implement the inference methods associated with such a result. More precisely, we provide the R code to perform inference on the smoothing distribution under dynamic probit models with four different methods:

1. i.i.d. sampling from the exact unified skew-normal distribution
2. expectation propagation (EP) approximation
3. partially factorized mean-field variational Bayes (PFM-VB) approximation
4. mean-field variational Bayes (MF-VB) approximation

Structure of the repository:

• the functions to implement the above methods can be found in the R source file Functions.R
• the financial dataset analyzed in the illustration can be found in Financial.Rdata, while the original entire dataset is publicly available at Yahoo Finance
• the code to reproduce the results in the paper is available at DynamicEP.R