Translator Disclaimer
2018 Detectability of nonparametric signals: higher criticism versus likelihood ratio
Marc Ditzhaus, Arnold Janssen
Electron. J. Statist. 12(2): 4094-4137 (2018). DOI: 10.1214/18-EJS1502


We study the signal detection problem in high dimensional noise data (possibly) containing rare and weak signals. Log-likelihood ratio (LLR) tests depend on unknown parameters, but they are needed to judge the quality of detection tests since they determine the detection regions. The popular Tukey’s higher criticism (HC) test was shown to achieve the same completely detectable region as the LLR test does for different (mainly) parametric models. We present a novel technique to prove this result for very general signal models, including even nonparametric $p$-value models. Moreover, we address the following questions which are still pending since the initial paper of Donoho and Jin: What happens on the border of the completely detectable region, the so-called detection boundary? Does HC keep its optimality there? In particular, we give a complete answer for the heteroscedastic normal mixture model. As a byproduct, we give some new insights about the LLR test’s behaviour on the detection boundary by discussing, among others, Pitmans’s asymptotic efficiency as an application of Le Cam’s theory.


Download Citation

Marc Ditzhaus. Arnold Janssen. "Detectability of nonparametric signals: higher criticism versus likelihood ratio." Electron. J. Statist. 12 (2) 4094 - 4137, 2018.


Received: 1 June 2018; Published: 2018
First available in Project Euclid: 13 December 2018

zbMATH: 07003238
MathSciNet: MR3890763
Digital Object Identifier: 10.1214/18-EJS1502

Primary: 62G10, 62G20
Secondary: 62G32


Vol.12 • No. 2 • 2018
Back to Top