For multivariate stationary time series many important properties, such as partial correlation, graphical models and autoregressive representations are encoded in the inverse of its spectral density matrix. This is not true for nonstationary time series, where the pertinent information lies in the inverse infinite dimensional covariance matrix operator associated with the multivariate time series. This necessitates the study of the covariance of a multivariate nonstationary time series and its relationship to its inverse. We show that if the rows/columns of the infinite dimensional covariance matrix decay at a certain rate then the rate (up to a factor) transfers to the rows/columns of the inverse covariance matrix. This is used to obtain a nonstationary autoregressive representation of the time series and a Baxter-type bound between the parameters of the autoregressive infinite representation and the corresponding finite autoregressive projection. The aforementioned results lay the foundation for the subsequent analysis of locally stationary time series. In particular, we show that smoothness properties on the covariance matrix transfer to (i) the inverse covariance (ii) the parameters of the vector autoregressive representation and (iii) the partial covariances. All results are set up in such a way that the constants involved depend only on the eigenvalue of the covariance matrix and can be applied in the high-dimensional settings with non-diverging eigenvalues.
more »
« less
Modeling Nonstationarity in Space and Time
We propose to model a spatio-temporal random field that has nonstationary covariance structure in both space and time domains by applying the concept of the dimension expansion method in Bornn et al. (2012). Simulations are conducted for both separable and nonseparable space-time covariance models, and the model is also illustrated with a streamflow dataset. Both simulation and data analyses show that modeling nonstationarity in both space and time can improve the predictive performance over stationary covariance models or models that are nonstationary in space but stationary in time.
more » « less- PAR ID:
- 10486019
- Publisher / Repository:
- Oxford University Press
- Date Published:
- Journal Name:
- Biometrics
- Volume:
- 73
- Issue:
- 3
- ISSN:
- 0006-341X
- Format(s):
- Medium: X Size: p. 759-768
- Size(s):
- p. 759-768
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Spatio-temporal data indexed by sampling locations and sampling time points are encountered in many scientific disciplines such as climatology, environ- mental sciences, and public health. Here, we propose a novel spatio-temporal expanding distance (STED) asymptotic framework for studying the proper- ties of statistical inference for nonstationary spatio-temporal models. In particular, to model spatio-temporal dependence, we develop a new class of locally stationary spatio-temporal covariance functions. The STED asymp- totic framework has a fixed spatio-temporal domain for spatio-temporal pro- cesses that are globally nonstationary in a rescaled fixed domain and locally stationary in a distance expanding domain. The utility of STED is illus- trated by establishing the asymptotic properties of the maximum likelihood estimation for a general class of spatio-temporal covariance functions. A simulation study suggests sound finite-sample properties and the method is applied to a sea-surface temperature dataset.more » « less
-
In many scientific fields, such as economics and neuroscience, we are often faced with nonstationary time series, and concerned with both finding causal relations and forecasting the values of variables of interest, both of which are particularly challenging in such nonstationary environments. In this paper, we study causal discovery and forecasting for nonstationary time series. By exploiting a particular type of state-space model to represent the processes, we show that nonstationarity helps to identify the causal structure, and that forecasting naturally benefits from learned causal knowledge. Specifically, we allow changes in both causal strengths and noise variances in the nonlinear state-space models, which, interestingly, renders both the causal structure and model parameters identifiable. Given the causal model, we treat forecasting as a problem in Bayesian inference in the causal model, which exploits the time-varying property of the data and adapts to new observations in a principled manner. Experimental results on synthetic and real-world data sets demonstrate the efficacy of the proposed methods.more » « less
-
more » « less
Abstract The standardized precipitation index (SPI) measures meteorological drought relative to historical climatology by normalizing accumulated precipitation. Longer record lengths improve parameter estimates, but these longer records may include signals of anthropogenic climate change and multidecadal natural climate fluctuations. Historically, climate nonstationarity has either been ignored or incorporated into the SPI using a quasi-stationary reference period, such as the WMO 30-yr period. This study introduces and evaluates a novel nonstationary SPI model based on Bayesian splines, designed to both improve parameter estimates for stationary climates and to explicitly incorporate nonstationarity. Using synthetically generated precipitation, this study directly compares the proposed Bayesian SPI model with existing SPI approaches based on maximum likelihood estimation for stationary and nonstationary climates. The proposed model not only reproduced the performance of existing SPI models but improved upon them in several key areas: reducing parameter uncertainty and noise, simultaneously modeling the likelihood of zero and positive precipitation, and capturing nonlinear trends and seasonal shifts across all parameters. Further, the fully Bayesian approach ensures all parameters have uncertainty estimates, including zero precipitation likelihood. The study notes that the zero precipitation parameter is too sensitive and could be improved in future iterations. The study concludes with an application of the proposed Bayesian nonstationary SPI model for nine gauges across a range of hydroclimate zones in the United States. Results of this experiment show that the model is stable and reproduces nonstationary patterns identified in prior studies, while also indicating new findings, particularly for the shape and zero precipitation parameters.
Significance Statement We typically measure how bad a drought is by comparing it with the historical record. With long-term changes in climate or other factors, however, a typical drought today may not have been typical in the recent past. The purpose of this study is to build a model that measures drought relative to a changing climate. Our results confirm that the model is accurate and captures previously noted climate change patterns—a drier western United States, a wetter eastern United States, earlier summer weather, and more extreme wet seasons. This is significant because this model can improve drought measurement and identify recent changes in drought.
-
more » « less
Abstract In complex systems with multiple variables monitored at high‐frequency, variables are not only temporally autocorrelated, but they may also be nonlinearly related or exhibit nonstationarity as the inputs or operation changes. One approach to handling such variables is to detrend them prior to monitoring and then apply control charts that assume independence and stationarity to the residuals. Monitoring controlled systems is even more challenging because the control strategy seeks to maintain variables at prespecified mean levels, and to compensate, correlations among variables may change, making monitoring the covariance essential. In this paper, a vector autoregressive model (VAR) is compared with a multivariate random forest (MRF) and a neural network (NN) for detrending multivariate time series prior to monitoring the covariance of the residuals using a multivariate exponentially weighted moving average (MEWMA) control chart. Machine learning models have an advantage when the data's structure is unknown or may change. We design a novel simulation study with nonlinear, nonstationary, and autocorrelated data to compare the different detrending models and subsequent covariance monitoring. The machine learning models have superior performance for nonlinear and strongly autocorrelated data and similar performance for linear data. An illustration with data from a reverse osmosis process is given.
