More Like this

1. Putting the Significance of Spectral Peaks on the Level: Implications for the 1470-Yr Peak in Greenland δ18O

   https://doi.org/10.1175/JCLI-D-22-0011.1  

   Huybers, Peter ( July 2022 , Journal of Climate)

   Spectral analysis of the Greenland Ice Sheet Project 2 (GISP2)δ¹⁸O record has been interpreted to show a 1/(1470 yr) spectral peak that is highly statistically significant (p< 0.01). The presence of such a peak, if accurate, provides an important clue about the mechanisms controlling glacial climate. As is standard, however, statistical significance was judged relative to a null model,H₀, consisting of an autoregressive order one process, AR(1). In this study,H₀is generalized using an autoregressive moving-average process, ARMA(p,q). A rule of thumb is proposed for evaluating the adequacy ofH₀involving comparing the expected and observed variances of the logarithm of a spectral estimate, which are generally consistent insomuch as removal of the ARMA structure from a time series results in an approximately level spectral estimate. An AR(1), or ARMA(1, 0), process is shown to be an inadequate representation of the GISP2δ¹⁸O structure, whereas higher-order ARMA processes result in approximately level spectral estimates. After suitably leveling GISP2δ¹⁸O and accounting for multiple hypothesis testing, multitaper spectral estimation indicates that the 1/(1470 yr) peak is insignificant. The seeming prominence of the 1/(1470 yr) peak is explained as the result of evaluating a spectrum involving higher-order ARMA structure and the peak having been selected on the basis of its seeming anomalous. The proposed technique for evaluating the significance of spectral peaks is also applicable to other geophysical records.

   A suitable null hypothesis is necessary for obtaining accurate test results, but a means for evaluating the adequacy of a null hypothesis for a spectral peak has been lacking. A generalized null model is presented in the form of an autoregressive, moving-average process whose adequacy can be gauged by comparing the observed and expected variance of log spectral density. Application of the method to the GISP2δ¹⁸O record indicates that spectral structure found at 1/(1470 yr) is statistically insignificant.

    

   more » « less
2. Autocovariance estimation in the presence of changepoints

   https://doi.org/10.1007/s42952-022-00173-5  

   Gallagher, Colin ; Killick, Rebecca ; Lund, Robert ; Shi, Xueheng ( June 2022 , Journal of the Korean Statistical Society)

   This article studies estimation of a stationary autocovariance structure in the presence of an unknown number of mean shifts. Here, a Yule–Walker moment estimator for the autoregressive parameters in a dependent time series contaminated by mean shift changepoints is proposed and studied. The estimator is based on first order differences of the series and is proven consistent and asymptotically normal when the number of changepoints m and the series length N satisfy 𝑚/𝑁→0 as 𝑁→∞. 

   more » « less
3. Maximum Likelihood Estimation of Linear Continuous Time Long Memory Processes with Discrete Time Data

   https://doi.org/10.1111/j.1467-9868.2005.00522.x  

   Tsai, Henghsiu ; Chan, K. S. ( November 2005 , Journal of the Royal Statistical Society Series B: Statistical Methodology)

   We develop a new class of time continuous autoregressive fractionally integrated moving average (CARFIMA) models which are useful for modelling regularly spaced and irregu-larly spaced discrete time long memory data. We derive the autocovariance function of a stationary CARFIMA model and study maximum likelihood estimation of a regression model with CARFIMA errors, based on discrete time data and via the innovations algorithm. It is shown that the maximum likelihood estimator is asymptotically normal, and its finite sample properties are studied through simulation. The efficacy of the approach proposed is demonstrated with a data set from an environmental study.

    

   more » « less
4. Fast, Sample-Efficient, Affine-Invariant Private Mean and Covariance Estimation for Subgaussian Distributions

   Brown, Gavin ; Hopkins, Samuel B. ; Smith, Adam ( July 2023 , 36th Annual Conference on Learning Theory (COLT 2023))

   We present a fast, differentially private algorithm for high-dimensional covariance-aware mean estimation with nearly optimal sample complexity. Only exponential-time estimators were previously known to achieve this guarantee. Given n samples from a (sub-)Gaussian distribution with unknown mean μ and covariance Σ, our (ε,δ)-differentially private estimator produces μ~ such that ∥μ−μ~∥Σ≤α as long as n≳dα2+dlog1/δ√αε+dlog1/δε. The Mahalanobis error metric ∥μ−μ^∥Σ measures the distance between μ^ and μ relative to Σ; it characterizes the error of the sample mean. Our algorithm runs in time O~(ndω−1+nd/ε), where ω<2.38 is the matrix multiplication exponent. We adapt an exponential-time approach of Brown, Gaboardi, Smith, Ullman, and Zakynthinou (2021), giving efficient variants of stable mean and covariance estimation subroutines that also improve the sample complexity to the nearly optimal bound above. Our stable covariance estimator can be turned to private covariance estimation for unrestricted subgaussian distributions. With n≳d3/2 samples, our estimate is accurate in spectral norm. This is the first such algorithm using n=o(d2) samples, answering an open question posed by Alabi et al. (2022). With n≳d2 samples, our estimate is accurate in Frobenius norm. This leads to a fast, nearly optimal algorithm for private learning of unrestricted Gaussian distributions in TV distance. Duchi, Haque, and Kuditipudi (2023) obtained similar results independently and concurrently. 

   more » « less
5. Fast, Sample-Efficient, Affine-Invariant Private Mean and Covariance Estimation for Subgaussian Distributions

   Brown, G. ; Hopkins, S. ; Smith, A. ( July 2023 , The Thirty Sixth Annual Conference on Learning Theory (COLT))

   We present a fast, differentially private algorithm for high-dimensional covariance-aware mean estimation with nearly optimal sample complexity. Only exponential-time estimators were previously known to achieve this guarantee. Given n samples from a (sub-)Gaussian distribution with unknown mean μ and covariance Σ, our (ϵ,δ)-differentially private estimator produces μ~ such that ∥μ−μ~∥Σ≤α as long as n≳dα2+dlog1/δ√αϵ+dlog1/δϵ. The Mahalanobis error metric ∥μ−μ^∥Σ measures the distance between μ^ and μ relative to Σ; it characterizes the error of the sample mean. Our algorithm runs in time O~(ndω−1+nd/\eps), where ω<2.38 is the matrix multiplication exponent.We adapt an exponential-time approach of Brown, Gaboardi, Smith, Ullman, and Zakynthinou (2021), giving efficient variants of stable mean and covariance estimation subroutines that also improve the sample complexity to the nearly optimal bound above.Our stable covariance estimator can be turned to private covariance estimation for unrestricted subgaussian distributions. With n≳d3/2 samples, our estimate is accurate in spectral norm. This is the first such algorithm using n=o(d2) samples, answering an open question posed by Alabi et al. (2022). With n≳d2 samples, our estimate is accurate in Frobenius norm. This leads to a fast, nearly optimal algorithm for private learning of unrestricted Gaussian distributions in TV distance.Duchi, Haque, and Kuditipudi (2023) obtained similar results independently and concurrently. 

   more » « less