bridge regression model with cauchy noise

We propose a bridge regression model when the noise follows cauchy distribution.we prove the identifiability of the proposed model using a bayesian representationof the model when the design matrix of the model is full rank. we specializeaic and bic as two criteria for regularization parameter selection based on theirdefinition for penalized models. in addition, we compare the performance of the proposedmodel with the usual lasso regression and quantile lasso regression usingsimulated and real datasets.