Readme.md

Setup python virtual environment and activate it

This assumes the python 2.7 Anaconda distribution is installed, pip package manager is available and you are on a linux machine

conda create -n fuzzy_bayes python=2.7
source activate fuzzy_bayes

Install the required libraries in the new environment

pip install -r requirements.txt

Run the following code to generate the data from a fuzzy inference system

This would also save a surface plot of the generated data, along with a .mat file for the data

python data_gen.py

Run MCMC code to estimate the parameters of the fuzzy system which generated the data

This would take a few minutes to finish.

python param_estimate.py

It would save the MCMC samples as a csv in a folder called 'run_1' This is just a quick demo with 5000 iterations of the MCMC chain, which is not enough for convergence. Change the maxsamples to 10,000 for a converged solution. The time taken would be a lot more.

Plot the posterior distribution of the membership functions and visualise

python post_processing.py