Sparse Vector Autoregressive Model Research Articles

Age-Coherent Mortality Modeling and Forecasting Using a Constrained Sparse Vector-Autoregressive Model

Accurate forecasts and analyses of mortality rates are essential to many practical issues, such as population projections and the design of pension schemes. Recent studies have considered a spatial–temporal autoregressive (STAR) model, in which the mortality rates of each age depend on their own historical values (temporality) and the neighboring cohort ages (spatiality). Despite the realization of age coherence and improved forecasting accuracy over the famous Lee-Carter (LC) model, the assumption of STAR that only the effects of the same and the neighboring cohorts exist can be too restrictive. In this study, we adopt a data-driven principle, as in a sparse vector autoregressive (SVAR) model, to improve the flexibility of the parametric structure of STAR and develop a constrained SVAR (CSVAR) model. To solve its objective function consisting of non-standard L2 and L1 penalties subject to constraints, we develop a new algorithm and prove the existence of the desirable age-coherence in CSVAR. Using empirical data from the United Kingdom, France, Italy, Spain, and Australia, we show that CSVAR consistently outperforms the LC, SVAR, and STAR models with respect to forecasting accuracy. The estimates and forecasts of the CSVAR model also demonstrate important demographic differences between these five countries.

High-dimensional structure learning of sparse vector autoregressive models using fractional marginal pseudo-likelihood

Learning vector autoregressive models from multivariate time series is conventionally approached through least squares or maximum likelihood estimation. These methods typically assume a fully connected model which provides no direct insight to the model structure and may lead to highly noisy estimates of the parameters. Because of these limitations, there has been an increasing interest towards methods that produce sparse estimates through penalized regression. However, such methods are computationally intensive and may become prohibitively time-consuming when the number of variables in the model increases. In this paper we adopt an approximate Bayesian approach to the learning problem by combining fractional marginal likelihood and pseudo-likelihood. We propose a novel method, PLVAR, that is both faster and produces more accurate estimates than the state-of-the-art methods based on penalized regression. We prove the consistency of the PLVAR estimator and demonstrate the attractive performance of the method on both simulated and real-world data.

Bootstrap based inference for sparse high-dimensional time series models

Fitting sparse models to high-dimensional time series is an important area of statistical inference. In this paper, we consider sparse vector autoregressive models and develop appropriate bootstrap methods to infer properties of such processes. Our bootstrap methodology generates pseudo time series using a model-based bootstrap procedure which involves an estimated, sparsified version of the underlying vector autoregressive model. Inference is performed using so-called de-sparsified or de-biased estimators of the autoregressive model parameters. We derive the asymptotic distribution of such estimators in the time series context and establish asymptotic validity of the bootstrap procedure proposed for estimation and, appropriately modified, for testing purposes. In particular, we focus on testing that large groups of autoregressive coefficients equal zero. Our theoretical results are complemented by simulations which investigate the finite sample performance of the bootstrap methodology proposed. A real-life data application is also presented.

Sparsity concepts and estimation procedures for high‐dimensional vector autoregressive models

High‐dimensional‐20 vector autoregressive (VAR) models are important tools for the analysis of multi‐variate time series. This article focuses on high‐dimensional time series and on the different regularized estimation procedures proposed for fitting sparse VAR models to such time series. Attention is paid to the different sparsity assumptions imposed on the VAR parameters and how these sparsity assumptions are related to the particular consistency properties of the estimators established. A sparsity scheme for high‐dimensional VAR models is proposed which is found to be more appropriate for the time series setting. Furthermore, it is shown that, under this sparsity setting, thresholding extends the consistency properties of regularized estimators to a wide range of matrix norms. Among other things, this enables application of the VAR parameters estimators to different problems, like forecasting or estimating the second‐order characteristics of the underlying VAR process. Extensive simulations compare the finite sample behavior of the different regularized estimators proposed using a variety of performance criteria.

Sparse Modeling for Cross-Industry Returns Predictability

We develop a new variational Bayes estimation method for large-dimensional sparse vector autoregressive models with exogenous predictors. Unlike existing Markov chain Monte Carlo (MCMC) and variational Bayes (VB) algorithms, our approach is not based on a structural form representation of the model, but directly shrinkage the lead-lag cross sectional interdependencies in the transition matrix. We compare the performance of our approach against state-of-the-art MCMC and VB methods, both in simulation and when forecasting a large cross section of industry portfolios spanning almost a hundred years of monthly data. The main results provide robust evidence that by directly shrinking weak cross sectional signals one can substantially improve both the statistical and economic out-of-sample performance of multivariate models for return predictability. This result holds across a variety of alternative shrinkage priors, such as Bayesian adaptive lasso, normal-gamma and horseshoe priors.

MORTALITY FORECASTING WITH A SPATIALLY PENALIZED SMOOTHED VAR MODEL

AbstractThis paper investigates a high-dimensional vector-autoregressive (VAR) model in mortality modeling and forecasting. We propose an extension of the sparse VAR (SVAR) model fitted on the log-mortality improvements, which we name “spatially penalized smoothed VAR” (SSVAR). By adaptively penalizing the coefficients based on the distances between ages, SSVAR not only allows a flexible data-driven sparsity structure of the coefficient matrix but simultaneously ensures interpretable coefficients including cohort effects. Moreover, by incorporating the smoothness penalties, divergence in forecast mortality rates of neighboring ages is largely reduced, compared with the existing SVAR model. A novel estimation approach that uses the accelerated proximal gradient algorithm is proposed to solve SSVAR efficiently. Similarly, we propose estimating the precision matrix of the residuals using a spatially penalized graphical Lasso to further study the dependency structure of the residuals. Using the UK and France population data, we demonstrate that the SSVAR model consistently outperforms the famous Lee–Carter, Hyndman–Ullah, and two VAR-type models in forecasting accuracy. Finally, we discuss the extension of the SSVAR model to multi-population mortality forecasting with an illustrative example that demonstrates its superiority in forecasting over existing approaches.

On constrained estimation of graphical time series models

Graphical time series models encode the conditional independence among the variables of a multivariate time series. An iterative method is proposed to estimate a graphical time series model based on a sparse vector autoregressive process. The method estimates both the autoregressive coefficients and the inverse of noise covariance matrix under sparsity constraints on both the coefficients and the inverse covariance matrix. This iterative method estimates a sparse vector autoregressive model by considering maximum likelihood estimation with the sparsity constraints as a biconcave problem, where the optimization problem becomes concave when either the autoregressive coefficients or the inverse noise covariance matrix is fixed. The method also imposes fewer restrictions in the estimation comparing to the use of a structural vector autoregressive model to study the dynamic interdependencies between time series variables.

Dimension reduction of frequency-based direct Granger causality measures on short time series

BackgroundThe mainstream in the estimation of effective brain connectivity relies on Granger causality measures in the frequency domain. If the measure is meant to capture direct causal effects accounting for the presence of other observed variables, as in multi-channel electroencephalograms (EEG), typically the fit of a vector autoregressive (VAR) model on the multivariate time series is required. For short time series of many variables, the estimation of VAR may not be stable requiring dimension reduction resulting in restricted or sparse VAR models. New methodThe restricted VAR obtained by the modified backward-in-time selection method (mBTS) is adapted to the generalized partial directed coherence (GPDC), termed restricted GPDC (RGPDC). Dimension reduction on other frequency based measures, such the direct directed transfer function (dDTF), is straightforward. ResultsFirst, a simulation study using linear stochastic multivariate systems is conducted and RGPDC is favorably compared to GPDC on short time series in terms of sensitivity and specificity. Then the two measures are tested for their ability to detect changes in brain connectivity during an epileptiform discharge (ED) from multi-channel scalp EEG. Comparison with existing method(s)It is shown that RGPDC identifies better than GPDC the connectivity structure of the simulated systems, as well as changes in the brain connectivity, and is less dependent on the free parameter of VAR order. ConclusionsThe proposed dimension reduction in frequency measures based on VAR constitutes an appropriate strategy to estimate reliably brain networks within short-time windows.

A Network Analysis of the Volatility of High Dimensional Financial Series

SummaryInterconnectedness between stocks and firms plays a crucial role in the volatility contagion phenomena that characterize financial crises, and graphs are a natural tool in their analysis. We propose graphical methods for an analysis of volatility interconnections in the Standard & Poor's 100 data set during the period 2000–2013, which contains the 2007–2008 Great Financial Crisis. The challenges are twofold: first, volatilities are not directly observed and must be extracted from time series of stock returns; second, the observed series, with about 100 stocks, is high dimensional, and curse-of-dimensionality problems are to be faced. To overcome this double challenge, we propose a dynamic factor model methodology, decomposing the panel into a factor-driven and an idiosyncratic component modelled as a sparse vector auto-regressive model. The inversion of this auto-regression, along with suitable identification constraints, produces networks in which, for a given horizon h, the weight associated with edge (i, j) represents the h-step-ahead forecast error variance of variable i accounted for by variable j's innovations. Then, we show how those graphs yield an assessment of how systemic each firm is. They also demonstrate the prominent role of financial firms as sources of contagion during the 2007–2008 crisis.

Sparse vector autoregressive modeling of audio signals and its application to the elimination of impulsive disturbances

Archive audio files are often corrupted by impulsive disturbances, such as clicks, pops and record scratches. This paper presents a new method for elimination of impulsive disturbances from stereo audio signals. The proposed approach is based on a sparse vector autoregressive signal model, made up of two components: one taking care of short-term signal correlations, and the other one taking care of long-term correlations. The method is evaluated on a set of clean audio signals contaminated with real click waveforms extracted from silent parts of old gramophone recordings.

Very-Short-Term Probabilistic Wind Power Forecasts by Sparse Vector Autoregression

A spatio-temporal method for producing very-short-term parametric probabilistic wind power forecasts at a large number of locations is presented. Smart grids containing tens, or hundreds, of wind generators require skilled very-short-term forecasts to operate effectively and spatial information is highly desirable. In addition, probabilistic forecasts are widely regarded as necessary for optimal power system management as they quantify the uncertainty associated with point forecasts. Here, we work within a parametric framework based on the logit-normal distribution and forecast its parameters. The location parameter for multiple wind farms is modeled as a vector-valued spatio-temporal process and the scale parameter is tracked by modified exponential smoothing. A state-of-the-art technique for fitting sparse vector autoregressive models is employed to model the location parameter and demonstrates numerical advantages over conventional vector autoregressive models. The proposed method is tested on a dataset of 5 min mean wind power generation at 22 wind farms in Australia. Five-min ahead forecasts are produced and evaluated in terms of point, and probabilistic forecast skill scores and calibration. Conventional autoregressive and vector autoregressive models serve as benchmarks.

희박 벡터자기상관회귀 모형을 이용한 한국의 미세먼지 분석

본 논문은 최근 많은 관심을 받는 미세먼지 (PM10)의 일별 평균농도에 대해서 전국 16개 시도에서 2008년부터 2011년까지 관측한 다변량 시계열 자료에 대한 연구이다. 다변량 시계열 모형을 이용해서 시간 및 공간에 대한 상관관계를 동시에 고려, 일변량 혹은 특정 지역에 국한해서 분석한 기존의 연구와 차별성을 두었다. 또한 Davis 등 (2013)이 제안한 부분 스펙트럼 일관성 (partial spectral coherence)을 통해 다른 지역간의 상호 의존성을 파악하고 이를 토대로 변수 선택을 통해 희박벡터자기회귀모형 (sVAR; sparse vector autoregressive model)을 적합하는 방법론을 적용하여 고차원 자료 분석의 단점 및 한계를 보완하였으며 예측력 비교를 통해서 sVAR 모형 적합의 타당성을 검증하였다. This paper considers multivariate time series modelling of PM10 data in Korea collected from 2008 to 2011. We consider both temporal and spatial dependencies of PM10 by applying the sparse vector autoregressive (sVAR) modelling proposed by Davis et al. (2013). It utilizes the partial spectral coherence to measure cross correlation between different regions, in turn provides the sparsity in the model while balancing the parsimony of model and the goodness of fit. It is also shown that sVAR performs better than usual vector autoregressive model (VAR) in forecasting.

Modeling gene expression regulatory networks with the sparse vector autoregressive model.

BackgroundTo understand the molecular mechanisms underlying important biological processes, a detailed description of the gene products networks involved is required. In order to define and understand such molecular networks, some statistical methods are proposed in the literature to estimate gene regulatory networks from time-series microarray data. However, several problems still need to be overcome. Firstly, information flow need to be inferred, in addition to the correlation between genes. Secondly, we usually try to identify large networks from a large number of genes (parameters) originating from a smaller number of microarray experiments (samples). Due to this situation, which is rather frequent in Bioinformatics, it is difficult to perform statistical tests using methods that model large gene-gene networks. In addition, most of the models are based on dimension reduction using clustering techniques, therefore, the resulting network is not a gene-gene network but a module-module network. Here, we present the Sparse Vector Autoregressive model as a solution to these problems.ResultsWe have applied the Sparse Vector Autoregressive model to estimate gene regulatory networks based on gene expression profiles obtained from time-series microarray experiments. Through extensive simulations, by applying the SVAR method to artificial regulatory networks, we show that SVAR can infer true positive edges even under conditions in which the number of samples is smaller than the number of genes. Moreover, it is possible to control for false positives, a significant advantage when compared to other methods described in the literature, which are based on ranks or score functions. By applying SVAR to actual HeLa cell cycle gene expression data, we were able to identify well known transcription factor targets.ConclusionThe proposed SVAR method is able to model gene regulatory networks in frequent situations in which the number of samples is lower than the number of genes, making it possible to naturally infer partial Granger causalities without any a priori information. In addition, we present a statistical test to control the false discovery rate, which was not previously possible using other gene regulatory network models.