Just a repository for understanding the basics of Bayesian linear regression by applying it to a simple polynomial fitting problem.
Clone the repository and install the dependencies using the python package manager poetry as
git clone https://github.com/ShantanuKodgirwar/bayesreg.git
cd bayesreg
poetry install --with dev --sync
This installs the package in a virtual environment and can be accessed with poetry shell
.
Implementations here mainly involve topics in Bayesian inference which are explained with some derivations under here. The following topics must also be reviewed,
- Markov chain Monte Carlo sampling to quantify the uncertainty of model parameters
- Sparsity prior, gradient-based priors
The following literature is considered in order of the level of complexity, mainly focusing on imaging.
- Bishop: Chapter 1 of "Pattern Recognition and Machine Learning"
- Ribes: Background on applications in imaging
- Hansen: Deconvolution and regularization
- MacKay: Bayesian perspective
- Minka: Bayesian linear regression