Bayesian Lasso Logistic Regression. Tibshirani and the Bayesian Lasso Speci cally the lasso estimate can be viewed as the mode of the posterior distribution of L argmax p jy2 when p j 2pexp jj jj 1 and the likelihood on pyj 2 NyjX 2I n. Bayesian Analysis 2010 5 Number 2 pp.
Another interpretation would be to use Laplace priors for the coefficients of your Bayesian model thereby making a sparsity assumption. Specifically the Bayesian Lasso. Penalized regression methods for simultaneous variable selection and coecient estimation especially those based on the lasso of Tibshirani 1996.
We describe an efficient Bayesian parallel GPU implementation of two classic statistical modelsthe Lasso and multinomial logistic regression.
Feb 07 2020 Bayesian logistic regression has the benefit that it gives us a posterior distribution rather than a single point estimate like in the classical also called frequentist approach. Any scheme L1 L2 Elasticnet would be great but Lasso is. The Bayesian Lasso estimates appear to be a compromise between the Lasso and ridge regression estimates. Our GPU implementations of Bayesian Lasso and multinomial logistic regression achieve 100-fold speedups on mid-level and high-end GPUs.
