1: \begin{abstract}

2:   We prove that kernel density estimation on symmetric spaces of non-compact type, whose $L_2$-risk was bounded above in previous work \cite{Asta-gKDE2}, in fact achieves a minimax rate of convergence.

3:   With this result, the story for kernel density estimation on all symmetric spaces is completed.  

4:   The idea in adapting the proof for Euclidean space is to suitably abstract vector space operations on Euclidean space to both actions of symmetric groups and reparametrizations of Helgason-Fourier transforms and to use the fact that the exponential map for symmetric spaces of non-compact type defines a diffeomorphism.     

5: %% Text of abstract

6: \end{abstract}

7: