Improving predictions with noisy functional data in linear models
Optimal estimation for Functional Linear Regression with Noisy Discretized Data
Machine Learning
Summary
When predicting a number based on data that comes as noisy curves measured at regular intervals, it can be tricky to get good estimates. The authors look at such a problem where you observe curves with noise and uneven precision. They first clean up these curves using a math technique involving Fourier methods, then estimate the relationship using a careful balance method to avoid overfitting. Their method comes with guarantees that it works well under reasonable conditions and does as well as possible when enough data points are available. They show their approach works with both simulated and real weather data.
functional linear regressionscalar-on-function regressionnoisy observationsFourier projectionpenalized least squaresoracle inequalitiesprediction erroreigenvalues decayminimax ratemodel dimension selection
Authors
Sixtine Sphabmixay
Abstract
In this paper, we consider the scalar-on-function linear regression model under a realistic sampling scheme in which the functional covariates are observed on a regular grid and contaminated by additive noise. We propose a two-step estimation procedure: first, the underlying curves are reconstructed from the discrete noisy observations using a Fourier-based projection method; second, the slope function is estimated by a penalized least-squares criterion over finite-dimensional trigonometric spaces, with data-driven selection of the model dimension. We establish oracle-type inequalities for the prediction error, both with respect to the reconstructed curves and to the true latent curves. Under regularity assumptions on the slope function and polynomial decay of the eigenvalues of the covariate, we derive convergence rates for the prediction error and show that our estimator attains the minimax rate when the number of grid points is sufficiently large. Finally, the proposed method is illustrated on simulated data and on a real meteorological dataset.