1: \begin{abstract}%   <- trailing '%' for backward compatibility of .sty file

2:  We consider shrinkage estimation of higher order Hilbert space-valued Bochner integrals in a  non-parametric setting. We propose estimators that shrink the $U$-statistic estimator of the Bochner integral towards a pre-specified target element in the Hilbert space. Depending on the degeneracy of the kernel of the $U$-statistic, we construct consistent shrinkage estimators 

3: %with fast rates of convergence, 

4: and develop oracle inequalities comparing the risks of the $U$-statistic estimator and its shrinkage version. Surprisingly, we show that the shrinkage estimator designed by assuming complete degeneracy of the kernel of the $U$-statistic is a consistent estimator even when the kernel is not complete degenerate. This work subsumes and improves upon \citet{muandet2016kernel} and \citet{zhou2019class}, which only handle mean element and covariance operator estimation in a reproducing kernel Hilbert space. We also specialize our results to normal mean estimation and show that for $d\ge 3$, the proposed estimator strictly improves upon the sample mean in terms of the mean squared error.

5:  %  all of existing cases in terms of oracle bounds. We design another shrinkage estimator  that obtains fast rates under assumption of complete degeneracy. As far as we know, ours is first work that obtains fast rates for shirnkage estimators under such assumptions. Surprisingly we show that estimator designed by assuming complete degeneracy is good estimator even when $r$ is not complete-degenerate and error term only gets increased by $\sqrt{n}.$ We also analyze estimator for mean of gaussian 

6: \end{abstract}

7: