1: \begin{abstract}

2: This paper deals with minimax rates of convergence for estimation of density functions on the real line. The densities are assumed to be location mixtures of normals, a global regularity requirement that creates subtle difficulties for the application of standard minimax lower bound methods.

3: Using novel Fourier and Hermite polynomial techniques, we determine the minimax optimal rate---slightly larger than the parametric rate---under squared error loss.

4: For Hellinger loss, we provide a minimax lower bound using ideas modified from the squared error loss case.

5: %Minimax rates of convergence for the problem of estimation in nonparametric models are often determined by strategically chosen submodels that capture the difficulty of the whole problem. The submodels are constructed to have separation properties, such as those required by Assouad's lemma, while still satisfying the constraints of the model. For the estimation of submodels, even more constraints are imported, which can make the calculation of minimax rates substantially more difficult. This paper illustrates the difficulty for a density estimation problem by means of a normal location mixture model, which actually has the same minimax rates as for a larger family of analytic densities. The analysis exploits a stability property of a two-parameter family of functions under convolution with the standard normal followed by Fourier transformation.

6: \end{abstract}

7: