Concentration inequalities and confidence bands for needlet density estimators on compact homogeneous manifolds

Abstract: Let

X_{1}, ..., X_{n}

be a random sample from some unknown probability density

f

defined on a compact homogeneous manifold

m a t h b f M

of dimension

d g e 1

. Consider a 'needlet frame'

p h i_{j e t a}

describing a localised projection onto the space of eigenfunctions of the Laplace operator on

m a t h b f M

with corresponding eigenvalues less than

2^{2 j}

, as constructed in cite{GP10}. We prove non-asymptotic concentration inequalities for the uniform deviations of the linear needlet density estimator

f_{n} (j)

obtained from an empirical estimate of the needlet projection

s u m_{e} t a p h i_{j e t a} i n t f p h i_{j e t a}

of

f

. We apply these results to construct risk-adaptive estimators and nonasymptotic confidence bands for the unknown density

f

. The confidence bands are adaptive over classes of differentiable and H"{older}-continuous functions on

m a t h b f M

that attain their H"{o}lder exponents.

Recommendations

Cites work

Cited in

(28)

This page was built for publication: Concentration inequalities and confidence bands for needlet density estimators on compact homogeneous manifolds