Short summary of the problem
I have a data contains lot of values close to zero. The data also have both positive and negative values. i tried gaussian family with identity link and zero inflated asymmetric laplace distribution, however, neither produces good posterior predictive check plots.
