Optimal difference-based variance estimators in time series: A general framework

Chan, Kin Wai

doi:10.1214/21-aos2154

Cited by 14 publications

(2 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The difference-based covariance estimators are good choices in the presence of change-points. The consistency could be reached under low-dimensional scenarios if the number of change-points is not too much and the change magnitudes are not too large (Rice, 1984;Chan, 2022). However, for asynchronous change pattern (1.1), where τ * j 's are not required to be the same, difference-based estimators may bring unnecessary bias accumulation, and thus may be not a good candidate, even under low-dimensional scenarios.…”

Section: Estimation Routines For the Covariance Matrixmentioning

confidence: 99%

Activation Discovery With FDR Control: Application to FMRI Data

Wen¹,

Wang²,

Zou³

et al. 2024

STAT SINICA

View full text Add to dashboard Cite

Time series from a large number of sources are ubiquitous, and some of them usually incur structural changes throughout data acquisition. For example, in fMRI analysis, some brain regions associated with task-related stimuli or in resting states become active. An activated time series can be composed of readings from an activated region. Of interest is to control the uncertainty of discovering time series in activation (viz., activated regions in fMRI analysis) via the false discovery rate (FDR) tool. We propose a simple yet effective method, incorporating unknown asynchronous change patterns and spatial dependence.Its validity in controlling the FDR is justified by asymptotic analysis. Numerical experiments indicate that the proposed method is both accurate and powerful.An implementation is provided in the R package SLIP.

show abstract

Section: Estimation Routines For the Covariance Matrixmentioning

confidence: 99%

Activation Discovery With FDR Control: Application to FMRI Data

Wen¹,

Wang²,

Zou³

et al. 2024

STAT SINICA

View full text Add to dashboard Cite

show abstract

“…We follow the suggestions discussed in Section 2.2 for choosing

g_{1}, \dots, g_{q}

and

q

. This type of kernel estimators has been well studied in the literature; see, e.g., Newey and West (1987b), Blackman and Tukey (1958), Andrews (1991), Andrews and Monahan (1992), Politis, and Romano (1994), White (2000), Chan and Yau (2017), Chan (2022b), and Chan (2022c). A popular choice of

K

is the Bartlett kernel

K (t) = (1 - | t |) 1 (| t | \leq 1)

.…”

Section: Inference In Coarsened Time Seriesmentioning

confidence: 99%

Inference in Coarsened Time Series via Generalized Method of Moments

Ip,

Chan

2024

Journal Time Series Analysis

View full text Add to dashboard Cite

We study statistical inference procedures in coarsened time series through the generalized method of moments. A new model for the coarsened time series via multiple potential outcomes is proposed. It can be naturally extended for inferring multi‐variate coarsened time series. We show that this framework generates a general class of estimators. It neatly generalizes the classical Horvitz–Thompson estimator for handling coarsened time series data. Asymptotic properties, including consistency and limiting distribution, of the proposed estimators are investigated. Estimators of the optimal weight matrix and the long‐run covariance matrix are also derived. In particular, confidence intervals of the mean function of the potential outcome as a function of coarsening index can be constructed. A real‐data application on air quality in the USA is investigated.

show abstract