Bayesian Elastic Net Regression

Dear Community,

I would like to use a Bayesian regression but something similar to the Elastic Net with l1 and l2 regularizations. Upon some research it seems changing prior to Gauss and Laplace distribution has similar effect to l1 and l2 regularization. I love to hear your thoughts on this matter.


Your research is pretty much right, did you have a specific question about implementation? Basically the standard deviation on the prior is going to control the degree of regularization (like the alpha or C hyperparameter in penalized regressions)