/Bayesian_Approximate_Inference

This project apply the Gibbs sampling and mean field methods to compute the inference and MAP inference

Primary LanguageMATLAB

Bayesian_Approximate_Inference

This project applies three methods to compute MAP inference and posterior inference

  1. the Gibbs sampling
  2. mean field methods to compute the inference and .
  3. In addition, the exact results are computed by the variable elimination method through Jupyter Notebook. ============================================================================== Code and report of ["Bayesian_Approximate_Inference"]

Prerequisites

The proposed method is implemented through Jupyter Notebook. The required packages include:

  • Matlab
  • Python 3
  • Jupyter Notebook

Getting started

  1. modify the path by the lcoation of the files in the folder of 'dataset';
  2. Run the function of 'Gibbs_sampling' 'mean_field' of the matlab codes;
  3. The 'Variable Elimination' method is in the Jupyter file of 'Proj1'

Bayesian Network

Gibbs ALgorithm

Mean Field algorithm

Performance