Testing independence of functional variables by angle covariance (Q2222234)

The authors consider the problem to test the independence of two functional random variables. Therefore, they propose a fully nonparametric test based on a novel dependence metric, which generalizes the projection covariance proposed for random vectors. They prove several desirable properties of the so-called angle covariance, including equivalence of its zero value and the independence of the two functional variables. Based on this, a V-statistic estimator of the angle covariance is constructed and it is shown that it has a Gaussian chaos limiting distribution under the independence null hypothesis and a normal limiting distribution under the alternative hypothesis. This statistic then gives rise to a consistent test. The authors furthermore suggest a random permutation method based implementation and investigate the behavior in simulations.