Extend available models and default priors in Bambi
- Mentors
- Ravin Kumar, Thomas Wiecki
- Organization
- NumFOCUS
This project aims to work on two related problems. On one hand, I would like to implement new family of generalized linear models, such as beta regression, robust linear regression (i.e. linear model with error following a T-Student distribution) as well as multinomial regression. On the other hand, it is also necessary to incorporate more automatic prior distribution alternatives that do not make use of the GLM module in statsmodels so new models can work smoothly without requiring the user to manually specify priors.