Optimal convergence rates for density estimation from grouped data

We derive the optimal convergence rates for density estimation based on aggregated observations under common smoothness conditions for symmetric densities. We study a procedure for data-driven bandwidth selection and give an extension to skew densities.