Supersmooth Testing on the Sphere over Analytic Classes

Peter S. Kim (1) , Ja-Yong Koo (2) , Thanh Mai Pham Ngoc (3)

1 Department of Mathematics and Statistics [Guelph]
2 Korea University
3 LM-Orsay - Laboratoire de Mathématiques d'Orsay

We consider the nonparametric goodness-of-fit test of the uniform density on the sphere when we have observations whose density is the convolution of an error density and the true underlying density. We will deal specifically with the supersmooth error case which includes the Gaussian distribution. Similar to deconvolution density estimation, the smoother the error density the harder is the rate recovery of the test problem. When considering nonparametric alternatives expressed over analytic classes, we show that it is possible to obtain original separation rates much faster than any logarithmic power of the sample size according to the ratio of the regularity index of the analytic class and the smoothness degree of the error. Furthermore, we show that our adaptive statistical procedure attains these optimal rates.

Mots clés: minimax hypothesis testing; Spherical deconvolution; adaptive testing; nonparametric alternatives; supersmooth error; analytic classes; rotational harmonics

Domaines: Statistiques [math.ST]

Peter S. Kim, Ja-Yong Koo, Thanh Mai Pham Ngoc. Supersmooth Testing on the Sphere over Analytic Classes. Journal of Nonparametric Statistics, 2016, 28 (1). ⟨hal-01082263v3⟩