The Permutation-Spectrum Test: Identifying Periodic Signals using the Maximum Fourier Intensity

TL;DR

This paper introduces a permutation-spectrum test that detects periodic signals in time-series data by analyzing the maximum Fourier intensity, offering robustness against non-normal noise and outperforming traditional methods.

Contribution

The paper proposes a new hypothesis test based on permutation methods to identify periodic signals, which is more robust than Fisher's spectrum test, especially with fat-tailed noise.

Findings

The test accurately detects periodic signals with a uniform p-value under the null hypothesis.

It is more robust than Fisher's spectrum test in the presence of fat-tailed noise.

Simulation results demonstrate improved detection power over existing methods.

Abstract

This paper examines the problem of testing whether a discrete time-series vector contains a periodic signal or is merely noise. To do this we examine the stochastic behaviour of the maximum intensity of the observed time-series vector and formulate a simple hypothesis test that rejects the null hypothesis of exchangeability if the maximum intensity spike in the Fourier domain is "too big" relative to its null distribution. This comparison is undertaken by simulating the null distribution of the maximum intensity using random permutations of the time-series vector. We show that this test has a p-value that is uniformly distributed for an exchangeable time-series vector, and that the p-value increases when there is a periodic signal present in the observed vector. We compare our test to Fisher's spectrum test, which assumes normality of the underlying noise terms. We show that our test is more robust than this test, and accommodates noise vectors with fat tails.