Multivariate Time Series Analysis Research Articles

Agricultural commodity futures prices prediction via long- and short-term time series network

ABSTRACTIn this study, we attempt to predict global agricultural commodity futures prices through analysis of multivariate time series. Our motivation is based on the notion that datasets of agricultural commodity futures prices involves a mixture of long- and short-term information, linear and non-linear structure, for which traditional approaches such as Auto-Regressive Integrated Moving Average (ARIMA) and Vector Auto-Regression (VAR) may fail. To tackle this issue, Long- and Short-Term Time-series Network (LSTNet) is applied for prediction. Empirical results show that LSTNet achieves better performance over that of several state-of-the-art baseline methods on average and in most periods based on three performance evaluation measures and two tests of performance difference.

Modeling Short Run Relationship between Naira-USD Exchange Rate and Crude Oil Price in Nigeria

The vector auto-regressive (VAR) model is one of the most successful, flexible, and easy to use models for the analysis of multivariate time series. It is a natural extension of the univariate auto-regressive model to a dynamic multivariate time series.The VAR model has proven to be especially useful for describing the dynamic behavior of economic and financial time series and for forecasting. It often provides superior forecasts to those from univariate time series models. The data used are monthlyobservations from January 2006 to October 2016 of Nigeria Crude Oil price and Naira to the dollar exchange rate. The VAR model was employed for modelling the data. The unit root test reveals that all the series are non-stationary at the level and stationary at first difference. The co-integration relations among the series indices were identified by applying Johansen’s cointegration test. The result of Johansen’s test indicates no existence of co-integration relation between the variables. The final result shows that a vector autoregressive (VAR) model of lag three with no co-integration equations best fits the data.

Frequent State Transition Patterns of Multivariate Time Series

Sequence pattern discovery is a key issue in multivariate time series analysis. Popular approaches first obtain the pattern of each single-variate time series and then obtain cross-variate associations. In this paper, we consider different variables at the same time during pattern construction, and propose a new type of pattern called State Transition pAttern with Periodic wildcard gaps (STAP). Compared to previous types, STAP reveals stronger cross associations among different variables and provides better interpretability for decision makers. We design an approach with two stages, namely frequent state discovery and pattern synthesis, to obtain frequent STAPs. We propose two pre-pruning and an Apriori-pruning techniques to speed up pattern discovery. We also propose two post-pruning techniques to simplify the output and a visualization way to support expert decision. Experimental results on four real-world datasets demonstrate 1) STAP captures the cross and temporal associations; 2) the five pruning and pattern synthesis techniques are quite effective; and 3) visualization technique greatly increases the readability of STAP.

Matrix factorization for multivariate time series analysis

Matrix factorization is a powerful data analysis tool. It has been used in multivariate time series analysis, leading to the decomposition of the series in a small set of latent factors. However, little is known on the statistical performances of matrix factorization for time series. In this paper, we extend the results known for matrix estimation in the i.i.d setting to time series. Moreover, we prove that when the series exhibit some additional structure like periodicity or smoothness, it is possible to improve on the classical rates of convergence.

A multivariate time series analysis of energy consumption, real output and pollutant emissions in a developing economy: New evidence from Nepal

Mountain economies will have to play a central role in attaining the global pursuit of green economic growth as crucial bearers of ecosystems goods and services. However, these economies are not adequately represented in the development policy debates in spite of their fundamental importance towards global sustainable development. This study examines the inter relationships between energy consumption, output and carbon emissions in a developing mountainous economy using an augmented Vector Autoregression model. Time-series data over the period 1975–2013 is studied applying a multivariate framework using population and gross fixed capital formation as additional variables for Nepal. Testing for Granger causality between integrated variables based on asymptotic theory reveals a long-run unidirectional Granger causality running from GDP to energy consumption, and a unidirectional Granger causality running from carbon emissions to GDP. We suggest that the government of Nepal can address energy poverty by accelerating the adoption energy conservation policies such as rationing energy consumption and energy efficiency improvements to narrow the energy supply-demand gap. The opportunity to promote the uptake of decentralised off-grid renewable technologies in remote areas and the large scale development of hydropower at the national level also needs to be prioritized. Our results remain robust across different estimators and contributes to an emerging literature on the nexus relationships between energy consumption, income and carbon emissions in mountainous developing economies.

A novel time-frequency multilayer network for multivariate time series analysis

Unveiling complex dynamics of natural systems from a multivariate time series represents a research hotspot in a broad variety of areas. We develop a novel multilayer network analysis framework, i.e. multivariate time-frequency multilayer network (MTFM network), to peer into the complex system dynamics. Through mapping the system features into different frequency-based layers and inferring interactions (edges) among different channels (nodes), the MTFM network allows efficiently integrating time, frequency and spatial information hidden in a multivariate time series. We employ two dynamic systems to illustrate the effectiveness of the MTFM network. We first apply the MTFM network to analyze the 48-channel measurements from industrial oil–water flows and reveal the complex dynamics ruling the transition of different flow patterns. The MTFM network is then utilized to analyze 30-channel fatigue driving electroencephalogram signals. The results demonstrate that MTFM network enables to quantitatively characterize brain behavior associated with fatigue driving. Our MTFM network enriches the multivariate time series analysis theory and helps to better understand the complicated dynamical behaviors underlying complex systems.

Dynamics of Health Agency Response and Public Engagement in Public Health Emergency: A Case Study of CDC Tweeting Patterns During the 2016 Zika Epidemic

BackgroundSocial media have been increasingly adopted by health agencies to disseminate information, interact with the public, and understand public opinion. Among them, the Centers for Disease Control and Prevention (CDC) is one of the first US government health agencies to adopt social media during health emergencies and crisis. It had been active on Twitter during the 2016 Zika epidemic that caused 5168 domestic noncongenital cases in the United States.ObjectiveThe aim of this study was to quantify the temporal variabilities in CDC’s tweeting activities throughout the Zika epidemic, public engagement defined as retweeting and replying, and Zika case counts. It then compares the patterns of these 3 datasets to identify possible discrepancy among domestic Zika case counts, CDC’s response on Twitter, and public engagement in this topic.MethodsAll of the CDC-initiated tweets published in 2016 with corresponding retweets and replies were collected from 67 CDC–associated Twitter accounts. Both univariate and multivariate time series analyses were performed in each quarter of 2016 for domestic Zika case counts, CDC tweeting activities, and public engagement in the CDC-initiated tweets.ResultsCDC sent out >84.0% (5130/6104) of its Zika tweets in the first quarter of 2016 when Zika case counts were low in the 50 US states and territories (only 560/5168, 10.8% cases and 662/38,885, 1.70% cases, respectively). While Zika case counts increased dramatically in the second and third quarters, CDC efforts on Twitter substantially decreased. The time series of public engagement in the CDC-initiated tweets generally differed among quarters and from that of original CDC tweets based on autoregressive integrated moving average model results. Both original CDC tweets and public engagement had the highest mutual information with Zika case counts in the second quarter. Furthermore, public engagement in the original CDC tweets was substantially correlated with and preceded actual Zika case counts.ConclusionsConsiderable discrepancies existed among CDC’s original tweets regarding Zika, public engagement in these tweets, and actual Zika epidemic. The patterns of these discrepancies also varied between different quarters in 2016. CDC was much more active in the early warning of Zika, especially in the first quarter of 2016. Public engagement in CDC’s original tweets served as a more prominent predictor of actual Zika epidemic than the number of CDC’s original tweets later in the year.

A fast LSH-based similarity search method for multivariate time series

Due to advances in mobile devices and sensors, there has been an increasing interest in the analysis of multivariate time series. Identifying similar time series is a core subroutine in many data mining and analysis problems. However, existing solutions mainly focus on univariate time series and fail to scale as the number of dimensions increase. Although, dimensionality reduction can reduce the impact of noisy information, the number of dimensions may still be too large. In this paper, an efficient approximation method is proposed based on locality sensitive hashing. It is a two-step solution which firstly retrieves candidate time series and then exploits their hash values to compute distance estimates for pruning. To probabilistically guarantee the result accuracy, an extensive error analysis has been conducted to determine appropriate LSH parameters. In addition, we also apply the proposed method to the PkNN classification and hierarchical clustering workloads. Finally, extensive experiments are conducted using both the real multivariate time series and the high-dimensional representations generated from univariate datasets in different query processing and data analysis workloads. Empirical results have verified the findings from the error analyses and demonstrated their benefits in terms of query efficiency when dealing with a collection of multivariate time series.

PASAR MODAL REGIONAL DALAM MASA KRISIS FINANSIAL 1997 DAN 2007: KAJIAN TERHADAP INTERDEPENDENSI BURSA EFEK ASIA TENGGARA

The devastating effect of the two world finansial crises had widely influenced not only on developed capital markets but also emerging ones, including the ASEAN regional markets. The crises have been commonly believed to have significant impact on the changing behaviour of the regional indices movements. This study investigates how the crises have affected the interrelation of stock indices’ movements amongst the five South East Asian countries. The multivariate time series analysis frameworks applied on series of the two sub-sample periods reveals the existing of a cointegrating relationship among the stock markets during the 1997 finansial crisis, but none of cointegrating vector to be found on the series of the 2007 crisis. The short run dynamic analyses conclude that the short run interrelation among the regional indices seems to be more intense during the 2007 finansial crisis period. For the latest period of crisis, the number of significant causal linkages between two variables on the series was greater than the other period. The analyses also show that the explanatory power of an endogenous variable to another in the system increased during the latest crisis, implying that the contagious effect of the crisis had increased the short run interdependence of the regional stock markets.

The impact of trait mindfulness on relational outcomes in novice yoga practitioners participating in an academic yoga course

Objective: This study examined the role of dispositional mindfulness in impacting the relational benefits of yoga in novice practitioners in an undergraduate academic yoga course. Participants: 21 college students enrolled in a yoga course throughout a 15-week academic semester (January–April 2016). Methods: Participants attended lab assessments at the beginning and end of the semester and were asked to complete eight consecutive Internet-based daily surveys across six separate bursts, yielding 48 data points for each participant. Results: Multivariate time-series analyses revealed within-person subgroup differences such that in general, those in the high trait mindfulness group gleaned benefits from yoga practice with respect to relational outcomes. Conclusions: This study demonstrated trait mindfulness impacts relational outcomes in novice yoga practitioners, suggesting those low in trait mindfulness may not garner immediate relational benefits from yoga and could benefit from additional strategies to bolster against any negative influences of initiating yoga practice.

Experimental study of interaction between liquid droplets in oil-in-water emulsions using multivariate time series analysis

Oil-in-water emulsions is frequently encountered in the industry of oil exploitation. The complex dynamic behavior of oil droplets brings about a great challenge in understanding its underlying flow regime. In the present study, we attempt to characterize flow instability of oil-in-water emulsions through measuring the information transfer between oil droplets. An experiment of vertical upward oil-in-water emulsions is conducted in a flow loop facility with 20 mm pipe inner diameter (ID), and multivariate time series are acquired by sampling the signals of a distributed sixteen-channel conductance mono-sensor probe array. The collected multivariate time series are firstly processed with an algorithm designated as transfer entropy (TE), through which the interaction between oil droplets at pipe cross section are investigated. Moreover, with the aid of multivariate weighted multiscale permutation entropy (MWMPE) algorithm, we also explore the global flow instability of oil droplets. Research results indicate that with the increasing distance from pipe center, the interaction between oil droplets gradually weakens. Besides, the increase in mixture velocity or the decrease in water-cut will intensify the interaction between oil droplets and global flow instability. Multivariate time series analysis can be considered as an effective method for characterizing flow instability in oil-in-water emulsions.

Motivation-Enhancing Human Resource Practices and the Dynamics of Organizational Performance

Scholars have long studied how the adoption of motivation-enhancing human resource (HR) practices, such as feedback and bonus pay, affects the level of organizational performance. Yet, today such practices are fairly widespread and so distinct questions emerge regarding ongoing performance dynamics after practice implementation. In this study, we ask how the routine activities that constitute implemented motivation- enhancing HR practices affect the dynamics of organizational performance. We hypothesize that individual instances of feedback and bonus pay generate positive but temporary deviations in performance, so that performance always returns to its status quo ante. In a multivariate time-series analysis applied to 169 weekly observations collected from an award-winning online-services firm, we uncover such a temporal dynamic. Moreover, we hypothesize and find that, in a context like ours, the effects of individual instances of feedback and bonus pay, while temporary, last longer than those of instances of online media activity (e.g., search-engine advertising). This study begins to address the need for research assessing the dynamic performance consequences of implemented HR practices, both separately and compared to other contextually relevant organizational practices.

Adaptive Bayesian Time–Frequency Analysis of Multivariate Time Series

ABSTRACTThis article introduces a nonparametric approach to multivariate time-varying power spectrum analysis. The procedure adaptively partitions a time series into an unknown number of approximately stationary segments, where some spectral components may remain unchanged across segments, allowing components to evolve differently over time. Local spectra within segments are fit through Whittle likelihood-based penalized spline models of modified Cholesky components, which provide flexible nonparametric estimates that preserve positive definite structures of spectral matrices. The approach is formulated in a Bayesian framework, in which the number and location of partitions are random, and relies on reversible jump Markov chain and Hamiltonian Monte Carlo methods that can adapt to the unknown number of segments and parameters. By averaging over the distribution of partitions, the approach can approximate both abrupt and slowly varying changes in spectral matrices. Empirical performance is evaluated in simulation studies and illustrated through analyses of electroencephalography during sleep and of the El Niño-Southern Oscillation. Supplementary materials for this article are available online.

Cross and joint ordinal partition transition networks for multivariate time series analysis

We propose to construct cross and joint ordinal pattern transition networks from multivariate time series for two coupled systems, where synchronizations are often present. In particular, we focus on phase synchronization, which is one of prototypical scenarios in dynamical systems. We systematically show that cross and joint ordinal patterns transition networks are sensitive to phase synchronization. Furthermore, we find that some particular missing ordinal patterns play crucial roles in forming the detailed structures in the parameter space whereas the calculations of permutation entropy measures often do not. We conclude that cross and joint ordinal partition transition network approaches provide complementary insights to the traditional symbolic analysis of synchronization transitions.

Online learning for vector autoregressive moving-average time series prediction

Abstract Multivariate time series analysis considers simultaneously multiple time series, which is much more complicated than the univariate time series analysis in general. VARMA (vector autoregressive moving-average) is one of the most mainstream multivariate time series model for time series prediction. However, the parameters of VARMA are often estimated in a batch manner in traditional multivariate statistical analysis and the noise terms are assumed Gaussian. In addition, the batch methods cannot perform satisfactorily in real-time prediction and the noise terms are unknown to us in real world. In this paper, we propose a novel online time series prediction framework for VARMA. We prove that VAR (vector autoregressive) can be used to mimic the underlying VARMA model under the online settings. Under this framework, we develop two effective algorithms VARMA-OGD and VARMA-ONS for this prediction problem assuming that the noise terms are generated stochastically and independently. The VARMA-OGD algorithm is based on the OGD (online gradient descent) algorithm, which is valid for general convex loss function. While the VARMA-ONS algorithm adopting the ONS (online newton step) algorithm is only valid for exp-concave loss function. Theoretical analysis shows that the regret bounds against the best VARMA prediction in hindsight of the proposed algorithms are sublinear of the number of samples. Furthermore, our experimental results further validate the effectiveness and robustness of our algorithms.

Nonuniform State Space Reconstruction for Multivariate Chaotic Time Series.

State space reconstruction is the foundation of chaotic system modeling. Selection of reconstructed variables is essential to the analysis and prediction of multivariate chaotic time series. As most existing state space reconstruction theorems deal with univariate time series, we have presented a novel nonuniform state space reconstruction method using information criterion for multivariate chaotic time series. We derived a new criterion based on low dimensional approximation of joint mutual information for time delay selection, which can be solved efficiently through the use of an intelligent optimization algorithm with low computation complexity. The embedding dimension is determined by conditional entropy, after which the reconstructed variables have relatively strong independence and low redundancy. The scheme, which integrates nonuniform embedding and feature selection, results in better reconstructions for multivariate chaotic systems. Moreover, the proposed nonuniform state space reconstruction method shows good performance in forecasting benchmark and actual multivariate chaotic time series.

Regularization and Interpolation of Positive Matrices

We construct certain matricial analogues of mass transport for positive-definite matrices of equal trace. The framework aims to devise ways of interpolating positive-definite matrices that tradeoff between “aligning up their eigenstructure” and “scaling the corresponding eigenvalues.” Motivation for the work is provided by power spectral analysis of multivariate time series where linear interpolation between matrix-valued power spectra generates push-pop unrealistic and undesirable artifacts.

Identification and functional characterization of HIV-associated neurocognitive disorders with large-scale Granger causality analysis on resting-state functional MRI.

Resting-state functional MRI (rs-fMRI), coupled with advanced multivariate time-series analysis methods such as Granger causality, is a promising tool for the development of novel functional connectivity biomarkers of neurologic and psychiatric disease. Recently large-scale Granger causality (lsGC) has been proposed as an alternative to conventional Granger causality (cGC) that extends the scope of robust Granger causal analyses to high-dimensional systems such as the human brain. In this study, lsGC and cGC were comparatively evaluated on their ability to capture neurologic damage associated with HIV-associated neurocognitive disorders (HAND). Functional brain network models were constructed from rs-fMRI data collected from a cohort of HIV+ and HIV - subjects. Graph theoretic properties of the resulting networks were then used to train a support vector machine (SVM) model to predict clinically relevant parameters, such as HIV status and neuropsychometric (NP) scores. For the HIV+ /- classification task, lsGC, which yielded a peak area under the receiver operating characteristic curve (AUC) of 0.83, significantly outperformed cGC, which yielded a peak AUC of 0.61, at all parameter settings tested. For the NP score regression task, lsGC, with a minimum mean squared error (MSE) of 0.75, significantly outperformed cGC, with a minimum MSE of 0.84 (p < 0.001, one-tailed paired t-test). These results show that, at optimal parameter settings, lsGC is better able to capture functional brain connectivity correlates of HAND than cGC. However, given the substantial variation in the performance of the two methods at different parameter settings, particularly for the regression task, improved parameter selection criteria are necessary and constitute an area for future research.

Job Seeker’s Allowance (JSA) benefit sanctions and labour market outcomes in Britain, 2001–2014

The dominant view among British policy-makers is that benefit sanctions for the unemployed who are claiming the Job Seeker’s Allowance (JSA) are effective at increasing flows from unemployment into sustainable employment. This paper tests this theory using aggregate cross-sectional data for Great Britain for the period May 2001 to December 2014. Descriptive analysis found the relationship between sanctions and labour market outcomes was ambiguous, while trends in labour market outcomes were highly correlated with labour market demand. Multivariate SVAR time-series analysis, controlling for labour market demand, found evidence that changes in the threat and use of sanctions had a positive impact on flows into work in the short run but not in the long term, and had no definitive impact on ILO unemployment at all. Interrupted time-series analyses suggest we cannot reject the null hypothesis that the impact of introduction of a new JSA sanctions regime in October 2012 (with higher financial penalties associated with being sanctioned) had no impact on flows into work from JSA. In Britain, intensifying the use of sanctions and introducing harsher penalties associated with being sanctioned has been largely ineffective at increasing flows from JSA into sustainable employment. Given the negative financial and social impacts of sanctions on those affected, and the lack of evidence of a sustained positive impact on employment, the basis for the new sanctions policy is unclear.

Vector autoregression and envelope model

Vector autoregression is an important technique for modelling multivariate time series and has been widely used in a variety of applications. Owing to its fast growth of parameters with the dimension of the time series vector, dimension reduction is often desirable in multivariate time series analysis. The envelope model is a new approach to achieve dimension reduction and allows efficient estimation in multivariate analysis. In this paper, we provide the first work to explore the application and extension of envelope models to multivariate time series data. We present the envelope and partial envelope formulations for vector autoregression and elaborate model selection, parameter estimation and asymptotic results. Simulations and real data analysis demonstrate the efficiency gains of the envelope vector autoregression models compared with the standard models in terms of estimation. Meanwhile, the envelope models can excel in prediction improvement.