Markov Chain Monte Carlo classification

Authors:

• Adrian Valente
• Armand Boschin
• Quentin Rebjock

This project is about the use of MCMC and simulated annealing for the Ising perceptron. Find details about the subject of the study here.

The report including performance and interpretation can be found here.

Details of the folders:

• doc : documentation about the project
• images : plots used for the report
• pickles : pickle files used during testing
• src : source code

Details of the code:

• utils.py functions implementing the model used in the rest of the files.
• markov_project.ipynb iPython notebook used for question 1
• Question2-3.py python script used for questions 2 and 3
• SA_crossval.py python script used for the tuning of the simulated annealing model
• SA_simulation.py python script used to try different values of alpha