DirichletLaplacePrior.coefficient_mean

DirichletLaplacePrior.coefficient_mean(context)

Zero everywhere: sparsity’s claim is that coefficients are zero.

There is no persistence hyperparameter here on purpose. Centering a selection prior away from zero would change what an exclusion means, so a caller who wants random-walk centering should difference the data or use the Minnesota family instead.

Return type:

NDArray[double]