Wavelet estimation of partially linear models

A difference-based approach in the partially linear model with dependent errors

We study asymptotic properties of estimators of parameter and non-parameter in a partially linear model in which errors are dependent. Using a difference-based and ordinary least square (DOLS) method, the estimator of an unknown parametric component is given and the asymptotic normality of the DOLS estimator is obtained. Meanwhile, the estimator of a nonparametric component is derived by the wavelet method, and asymptotic normality and the weak convergence rate of the wavelet estimator are discussed. Finally, the performance of the proposed estimator is evaluated by a simulation study.

Variable selection in partially linear wavelet models

Variable selection is fundamental in high-dimensional statistical modelling, including non-and semiparametric regression. However, little work has been done for variable selection in a partially linear model (PLM). We propose and study a unified approach via double penalized least squares, retaining good features of both variable selection and model estimation in the framework of PLM. The proposed method is distinguished from others in that the penalty functions combine the l1 penalty coming from wavelet thresholding in the non-parametric component with the l1 penalty from the lasso in the parametric component. Simulations are used to investigate the performances of the proposed estimator in various settings, illustrating its effectiveness for simultaneous variable selection as well as estimation.

Nonparametric trend estimation in the presence of fractal noise: application to fMRI time-series analysis

Unknown low frequency fluctuations called "trend" are observed in noisy time-series measured for different applications. In some disciplines, they carry primary information while in other fields such as functional magnetic resonance imaging (fMRI) they carry nuisance effects. In all cases, however, it is necessary to estimate them accurately. In this paper, a method for estimating trend in the presence of fractal noise is proposed and applied to fMRI time-series. To this end, a partly linear model (PLM) is fitted to each time-series. The parametric and nonparametric parts of PLM are considered as contributions of hemodynamic response and trend, respectively. Using the whitening property of wavelet transform, the unknown components of the model are estimated in the wavelet domain. The results of the proposed method are compared to those of other parametric trend-removal approaches such as spline and polynomial models. It is shown that the proposed method improves activation detection and decreases variance of the estimated parameters relative to the other methods.