- This repository is for the Machine Learning for Probabilistic Programming Project.
- We focus on the criticism aspect of the Box's loop, and will analyze criticism frameworks such as the Posterior Predictive Check (PPC) and Population Predictive Check (POP-PC), and test them on different types of probabilistic models that were not mentioned in the original paper.
.
├── data # store data here
├── mlpp_pop_pc # Github source code here
│ ├── final-project
│ ├── requirements.txt
└── └── README.md
Create a Python 3 (Preferably 3.7) virtual environment (make sure you have Python 3 in /usr/bin/)
python3 -m venv venvthen
source venv/bin/activate
pip install ipykernel
python -m ipykernel install --user --name venv --display-name "CHOOSENAME"Install requirements via pip
pip install -r requirements.txtjupyter notebook- Since the data is too big to place on Github, we have placed the required data on a Google Drive in the following link: https://drive.google.com/open?id=1yBBesBV5kQj5aPwDz4hq6uu6uJ9AgAAM
Place the data as specified in the directory.
- Ranganath & Blei 2019 Population Predictive Checks
- Salakhutdinov & Mnih 2008 Probabilistic Matrix Factorization
- Salakhutdinov & Mnih 2008 Bayesian Probabilistic Matrix Factorization using Markov Chain Monte Carlo
- Gopalan & Hofman & Blei 2013 Scalable Recommendation with Poisson Factorization
- Gelman et al. 2013 Bayesian Data Analysis
- Barnard & McCulloch & Meng 2000 Modeling Covariance Matrices in Terms of Standard Deviations and Correlations, with Application to Shrinkage
PEP8 guidelines: https://realpython.com/python-pep8/