Recovering covariance from functional fragments

We consider nonparametric estimation of a covariance function on the unit square, given a sample of discretely observed fragments of functional data. When each sample path is observed only on a subinterval of length , one has no statistical information on the unknown covariance outside a -band around the diagonal. The problem seems unidentifiable without parametric assumptions, but we show that nonparametric estimation is feasible under suitable smoothness and rank conditions on the unknown covariance. This remains true even when the observations are discrete, and we give precise deterministic conditions on how fine the observation grid needs to be relative to the rank and fragment length for identifiability to hold true. We show that our conditions translate the estimation problem to a low-rank matrix completion problem, construct a nonparametric estimator in this vein, and study its asymptotic properties. We illustrate the numerical performance of our method on real and simulated data.

Functional data analysis by matrix completion

Marie-Hélène Descary

Traditional approaches to analysing functional data typically follow a two-step procedure, consisting in first smoothing and then carrying out a functional principal component analysis. The idea underlying this procedure is that functional data are well approximated by smooth functions, and that rough variations are due to noise. However, it may very well happen that localised features are rough at a global scale but still smooth at some finer scale. In this thesis we put forward a new statistical approach for functional data arising as the sum of two uncorrelated components: one smooth plus one rough. We give non-parametric conditions under which the covariance operators of the smooth and of the rough components are jointly identifiable on the basis of discretely observed data: the covariance operator corresponding to the smooth component must be of finite rank and have real analytic eigenfunctions, while the one corresponding to the rough component must have a banded covariance function. We construct consistent estimators of both covariance operators without assuming knowledge of the true rank or bandwidth. We then use them to estimate the best linear predictors of the the smooth and the rough components of each functional datum. In both the identifiability and the inference part, we do not follow the usual strategy used in functional data analysis which is to first employ smoothing and work with continuous estimate of the covariance operator. Instead, we work directly with the covariance matrix of the discretely observed data, which allows us to use results and tools from linear algebra. In fact, we show that the whole problem of uniquely recovering the covariance operator of the smooth component given the one of the raw data can be seen as a low-rank matrix completion problem, and we make great use of a classical relation between the rank and the minors of a matrix to solve this matrix completion problem. The finite-sample performance of our approach is studied by means of simulation study.

EPFL2017

Dispersion operators and resistant second-order functional data analysis

David Kraus, Victor Panaretos

Inferences related to the second-order properties of functional data, as expressed by covariance structure, can become unreliable when the data are non-Gaussian or contain unusual observations. In the functional setting, it is often difficult to identify atypical observations, as their distinguishing characteristics can be manifold but subtle. In this paper, we introduce the notion of a dispersion operator, investigate its use in probing the second-order structure of functional data, and develop a test for comparing the second-order characteristics of two functional samples that is resistant to atypical observations and departures from normality. The proposed test is a regularized M-test based on a spectrally truncated version of the Hilbert-Schmidt norm of a score operator defined via the dispersion operator. We derive the asymptotic distribution of the test statistic, investigate the behaviour of the test in a simulation study and illustrate the method on a structural biology dataset.

Oxford University Press2012

Hybrid regularisation and the (in)admissibility of ridge regression in infinite dimensional Hilbert spaces

Anirvan Chakraborty, Victor Panaretos

We consider the problem of estimating the slope function in a functional regression with a scalar response and a functional covariate. This central problem of functional data analysis is well known to be ill-posed, thus requiring a regularised estimation procedure. The two most commonly used approaches are based on spectral truncation or Tikhonov regularisation of the empirical covariance operator. In principle, Tikhonov regularisation is the more canonical choice. Compared to spectral truncation, it is robust to eigenvalue ties, while it attains the optimal minimax rate of convergence in the mean squared sense, and not just in a concentration probability sense. In this paper, we show that, surprisingly, one can strictly improve upon the performance of the Tikhonov estimator in finite samples by means of a linear estimator, while retaining its stability and asymptotic properties by combining it with a form of spectral truncation. Specifically, we construct an estimator that additively decomposes the functional covariate by projecting it onto two orthogonal subspaces defined via functional PCA; it then applies Tikhonov regularisation to the one component, while leaving the other component unregularised. We prove that when the covariate is Gaussian, this hybrid estimator uniformly improves upon the MSE of the Tikhonov estimator in a non-asymptotic sense, effectively rendering it inadmissible. This domination is shown to also persist under discrete observation of the covariate function. The hybrid estimator is linear, straightforward to construct in practice, and with no computational overhead relative to the standard regularisation methods. By means of simulation, it is shown to furnish sizeable gains even for modest sample sizes.

2019

Functional data analysis by matrix completion

Marie-Hélène Descary

EPFL2017