Search

Video stabilization is the process of improving the video quality by removing annoying fluctuant motion caused by camera jittering. A key issue of a successful solution is the temporal adaptability to motion and the overall robustness with respect to different motion types. However, most previous methods usually produce non-motion adaptive stabilized videos. In other words, under-smoothing in slow motion segments and over-smoothing in rapid motion segments will be produced for complex shaky videos. To overcome these drawbacks, we propose a novel video stabilization approach using a motion morphological component (MMC decomposition. Specifically, the observed motion is decomposed into three MMCs: low-frequency smoothed (LFS motion, high-frequency compensatory (HFC motion, and shaky motion. LFS motion helps to largely stabilize videos, and HFC motion helps to recover missing motion to deal with over-smoothing. Subsequently, we present an MMC-based model to retrieve the desired smoothed motion, in which weighted nuclear norm and autoregression priors are used for LFS motion, while a sparsity prior is adopted for HFC motion. In addition, we design an adaptive weight setting scheme to detect rapid motions and to calculate the optimal weights. Finally, we develop a stabilization algorithm under the Alternating Direction Method of Multipliers (ADMM framework. Experimental results demonstrate that our method can achieve high-quality results compared with that of other state-of-the-art stabilization methods in terms of robustness and efficiency, both quantitatively and qualitatively.

Monitoring economic conditions in real time, or nowcasting, and Big Data analytics share some challenges, sometimes called the three “Vs”. Indeed, nowcasting is characterized by the use of a large number of time series (Volume), the complexity of the data covering various sectors of the economy, with different frequencies and precision and asynchronous release dates (Variety), and the need to incorporate new information continuously and in a timely manner (Velocity). In this paper, we explore three alternative routes to nowcasting with Bayesian Vector Autoregressive (BVAR) models and find that they can effectively handle the three Vs by producing, in real time, accurate probabilistic predictions of US economic activity and a meaningful narrative by means of scenario analysis.

This work introduces a methodology to estimate coffee prices based on the use of Extreme Learning Machines. The process is initiated by identifying the presence of nonstationary components, like seasonality and trend. These components are withdrawn if they are found. Next, the temporal lags are selected based on the response of the Partial Autocorrelation Function filter. As predictors, we address the following models: Exponential Smoothing (ES), Autoregressive (AR) and Autoregressive Integrated and Moving Average (ARIMA) models, Multilayer Perceptron (MLP) and Extreme Learning Machines (ELMs) neural networks. The computational results based on three error metrics and two coffee types (Arabica and Robusta) showed that the neural networks, especially the ELM, can reach higher performance levels than the other models. The methodology, which presents preprocessing stages, lag selection, and use of ELM, is a novelty that contributes to the coffee prices forecasting field.

This paper considers a problem of estimating brain effective connectivity from EEG signals using a Granger causality (GC) concept characterized on state-space models. We propose a state-space model for explaining coupled dynamics of the source and EEG signals where EEG is a linear combination of sources according to the characteristics of volume conduction. Our formulation has a sparsity prior on the source output matrix that can further classify active and inactive sources. The scheme is comprised of two main steps: model estimation and model inference to estimate brain connectivity. The model estimation consists of performing a subspace identification and the active source selection based on a group-norm regularized least-squares. The model inference relies on the concept of state-space GC that requires solving a discrete-time Riccati equation for the covariance of estimation error. We verify the performance on simulated data sets that represent realistic human brain activities under several conditions including percentages of active sources, a number of EEG electrodes and the location of active sources. The performance of estimating brain networks is compared with a two-stage approach using source reconstruction algorithms and VAR-based Granger analysis. Our method achieved better performances than the two-stage approach under the assumptions that the true source dynamics are sparse and generated from state-space models. The method is applied to a real EEG SSVEP data set and we found that the temporal lobe played a role of a mediator of connections between temporal and occipital areas, which agreed with findings in previous studies.

In this paper, we investigate brain directed connectivity (BDC) networks for emotion recognition using electroencephalogram (EEG) source signals that were estimated from high-density sensor EEG signals, for the first time. Currently, a variety of features extracted from sensor EEG signals are used for emotion recognition. However, they cannot unambiguously describe the location of emotions associated with neural activities and information propagation or the interaction between brain regions. In addition, most current studies use low-density sensor EEG signals. Moreover, source signals estimated from high-density sensor EEG signal have not been employed for emotion recognition to date. We designed a BDC network-based framework using EEG source signals to investigate emotion recognition. The global cortex factor-based multivariate autoregressive (GCF-MVAR) method was utilized to extract emotion-related BDC features. Our study revealed that the combined BDC and DE features facilitated a recognition accuracy of up to 89.58 %, which is higher than the rate obtained from BDC features and DE features alone. The sensor features derived from high-density EEG signals also exhibited higher recognition accuracy compared to low-density EEG signals. These findings suggest that BDC features derived from EEG source signals can better characterize human emotional states and are meaningful for emotion recognition.

Long horizon state of health (SOH) monitoring and remaining useful life (RUL) prediction are of industrial value in prognostic and health management (PHM) of lithium-ion batteries (LIBs) to ensure their reliable functionality by early detection. Machine Learning, as a data-driven health diagnostic technique, has been widely utilized in solitary and hybrid structures. However, an accurate SOH estimation and RUL prediction method with less computational burden are highly desirable for the online state prediction in an electric vehicle application. This paper evaluates nonlinear autoregressive with external input (NARX) recurrent neural network (RNN) and time delay neural network (TDNN) in their prediction precision using the NASA dataset. The superior method, NARXRNN, is employed for two different datasets to estimate the battery's SOH and predict its RUL on a broad horizon. The results reveal the outstanding performance by presenting the root mean square error within 3% and mean absolute error within 2% for unseen data. Therefore, this method is capable to accurately predict the SOH of LIBS from historical data at low computational complexity. It is a promising model for long horizon SOH and RUL prediction and practical for online applications.

The brain must extract behaviorally relevant latent variables from the signals streamed by the sensory organs. Such latent variables are often encoded in the dynamics that generated the signal rather than in the specific realization of the waveform. Therefore, one problem faced by the brain is to segment time series based on underlying dynamics. We present two algorithms for performing this segmentation task that are biologically plausible, which we define as acting in a streaming setting and all learning rules being local. One algorithm is model-based and can be derived from an optimization problem involving a mixture of autoregressive processes. This algorithm relies on feedback in the form of a prediction error, and can also be used for forecasting future samples. In some brain regions, such as the retina, the feedback connections necessary to use the prediction error for learning are absent. For this case, we propose a second, model-free algorithm that uses a running estimate of the autocorrelation structure of the signal to perform the segmentation. We show that both algorithms do well when tasked with segmenting signals drawn from autoregressive models with piecewise-constant parameters. In particular, the segmentation accuracy is similar to that obtained from oracle-like methods in which the ground-truth parameters of the autoregressive models are known. We also test our methods on datasets generated by alternating snippets of voice recordings. We provide implementations of our algorithms at https://github.com/ttesileanu/bio-time-series., Comment: v2.1; 34 pages, 14 figures

Introduction Multivariate time series prediction of infectious diseases is significant to public health, and the deep learning method has attracted increasing attention in this research field. Material and methods An adaptively temporal graph convolution (ATGCN) model, which learns the contact patterns of multiple age groups in a graph-based approach, was proposed for COVID- 19 and influenza prediction. We compared ATGCN with autoregressive models, deep sequence learning models, and experience- based ATGCN models in short-term and long-term prediction tasks. Results Results showed that the ATGCN model performed better than the autoregressive models and the deep sequence learning models on two datasets in both short-term (12.5% and 10% improvements on RMSE) and long-term (12.4% and 5% improvements on RMSE) prediction tasks. And the RMSE of ATGCN predictions fluctuated least in different age groups of COVID- 19 (0.029 ± 0.003) and influenza (0.059±0.008). Compared with the Ones-ATGCN model or the Pre-ATGCN model, the ATGCN model was more robust in performance, with RMSE of 0.0293 and 0.06 on two datasets when horizon is one. Discussion Our research indicates a broad application prospect of deep learning in the field of infectious disease prediction. Transmission characteristics and domain knowledge of infectious diseases should be further applied to the design of deep learning models and feature selection. Conclusions The ATGCN model addressed the multivariate time series forecasting in a graph-based deep learning approach and achieved robust prediction on the confirmed cases of multiple age groups, indicating its great potentials for exploring the implicit interactions of multivariate variables.

The paper presents a prediction of non-linear exogenous signal by optimized intelligent auto-regressive neural network model (ARNN). A signal comprises of two sets of data called deterministic and error. The former type of data represents the degradation index of a signal, while the error is the uncertainties associated with the signal. To understand and predict signals, a intelligent approach is taken through the use of ARNN model. In this approach, the rst step is to diagnose whether a time series signal is normally distributed or not by utilizing the Jarque-Bera test. The high and low volatility data ele- ments can be separated via kurtosis hypothesis. The deterministic component of the signal is also predicted by developing a neural network based non-linear autoregressive model (NN-NARX) and the error component by using a linear model. The nal forecast is formed by combining the results determined from each of the models and evaluated using the mean square error results. Vali- dation of the prediction is obtained through a comparison of the results with other models such as ARNN, traditional ARMX, and NARX models. The re- sults show that the proposed model provides improved predictions, minimize high dependence on design parameters with low computational cost.

Recently, many efforts have been devoted to speeding up neural machine translation models. Among them, the non-autoregressive translation (NAT) model is promising because it removes the sequential dependence on the previously generated tokens and parallelizes the generation process of the entire sequence. On the other hand, the autoregressive translation (AT) model in general achieves a higher translation accuracy than the NAT counterpart. Therefore, a natural idea is to fuse the AT and NAT models to seek a trade-off between inference speed and translation quality. This paper proposes an ARF-NAT model (NAT with auxiliary representation fusion) to introduce the merit of a shallow AT model to an NAT model. Three functions are designed to fuse the auxiliary representation into the decoder of the NAT model. Experimental results show that ARF-NAT outperforms the NAT baseline by 5.26 BLEU scores on the WMT’14 German-English task with a significant speedup (7.58 times) over several strong AT baselines.

In this paper we present a thorough description of the digital twin development for a continuous pharmaceutical powder blending process in accordance with the Process Analytical Technologies (PAT) and Quality by Design (QbD) guidelines. A low-dosage system of caffeine active pharmaceutical ingredient (API) and dextrose excipient was examined via continuous blending experiments. Near infrared (NIR) spectroscopy-based process analytics were applied; quantitative evaluation of spectra was achieved using multivariate data analysis. The blending system was represented with mechanistic residence time distribution (RTD) models and two types of recurrent artificial neural networks (ANN), experimental datasets were used for model training or fitting and validation. Detailed comparison of the two modelling approaches, the optimization of the model-based digital twin, and efficiency of the soft sensor-based process monitoring is presented through several validating simulations. Both RTD models and nonlinear autoregressive neural networks demonstrated excellent predictive power for the low dosage blending process. RTD models can prove to be more advantageous in industrial development as they are less resource-intensive to develop and prediction accuracy on low concentration levels lacks dependency from the precision of chemometric calibration. Reduced material costs and limited development timeframe render the digital twin an efficient tool in technological development.

Wind speed predictions are essential for wind power management and wind farm operation. However, due to the high volatility and nonstationarity of measured wind data, it is often difficult to achieve an accurate prediction. This study proposes a hybrid approach that consists of two stages, i.e., data preprocessing and wind speed predicting, to improve the accuracy of short-term wind speed prediction. A preprocessing algorithm for the transformation and standardization of hourly mean wind speed is utilized to remove the non-Gaussian distribution of wind data and diurnal nonstationarity. Several statistical models and artificial intelligence models are then adopted in the second stage of the prediction process, including a persistence model, autoregressive model, autoregressive moving average model and backpropagation neural network. The proposed approach is developed based on the weighted averaging of these models and error optimization theory. Finally, wind speed data for 12 months from two meteorological towers located in Yanan, China, are investigated to demonstrate the effectiveness and accuracy of the proposed approach for multistep wind speed predictions, and its performance is then compared with several existing prediction models. The results indicate that the prediction accuracy improves significantly after preprocessing with the proposed approach, outperforming all the existing aforementioned models.

Many previous studies have developed decomposition and ensemble models to improve runoff forecasting performance. However, these decomposition-based models usually introduce large decomposition errors into the modeling process. Since the variation in runoff time series is greatly driven by climate change, many previous studies considering climate change focused on only rainfall-runoff modeling, with few meteorological factors as input. Therefore, a climate-driven streamflow forecasting (CDSF) framework was proposed to improve the runoff forecasting accuracy. This framework is realized by using principal component analysis (PCA), long short-term memory (LSTM) and Bayesian optimization (BO), referred to as PCA-LSTM-BO. To validate the effectiveness and superiority of the PCA-LSTM-BO method along with one autoregressive LSTM model and two other CDSF models based on PCA, BO, and either support vector regression (SVR) or gradient boosting regression trees (GBRT), namely, PCA-SVR-BO and PCA-GBRT-BO, respectively, were compared. A generalization performance index based on the Nash-Sutcliffe efficiency (NSE), called the GI(NSE) value, is proposed to evaluate the generalizability of the model. The results show that (1) the proposed model is significantly better than the other benchmark models in terms of the mean square error (MSE =0.819, and GI(NSE)

PurposeThis study focuses on accessing the impact of lockdown implemented to curb the pandemic of coronavirus disease 2019 (COVID-19) on prices of potato and onion crops using the time series analysis techniques.Design/methodology/approachThe present study uses secondary price series data for both crops. Along with the study of percent increase or decrease, the time series analysis techniques of autoregressive integrated moving average (ARIMA) and generalized autoregressive conditional heteroskedasticity (GARCH), as well as machine learning; neural network autoregressive (NNAR) models were used to model the prices. For the purpose of comparison, the data from past years were taken as the period of normalcy. The behaviour of the forecasts for the normal periods and during the pandemic based on respective datasets was compared.FindingsThe results show that there was an unprecedented rise in prices during the months of lockdown. It could be attributed to the decline in arrivals due to several reasons like issues with transportation and labour availability. Also, towards the end of lockdown (May 2020), the prices seemed to decrease. Such a drop could be attributed to the relaxations in lockdown and reduced demand. The study also discusses that how some unique approaches like e-marketing, localized resource development for attaining self-sufficiency and developing transport chain, especially, for agriculture could help in such a situation of emergency.Research limitations/implicationsA more extensive study could be conducted to mark the factors specifically that caused the increase in price.Originality/valueThe study clearly marks that the prices of the crops increased more than expectations using time series methods. Also, it surveys the prevailing situation through available resources to link up the reasons behind it.

Recently, surface electromyography (sEMG) emerged as a novel biometric authentication method. Since EMG system parameters, such as the feature extraction methods and the number of channels, have been known to affect system performances, it is important to investigate these effects on the performance of the sEMG-based biometric system to determine optimal system parameters. In this study, three robust feature extraction methods, Time-domain (TD) feature, Frequency Division Technique (FDT), and Autoregressive (AR) feature, and their combinations were investigated while the number of channels varying from one to eight. For these system parameters, the performance of sixteen static wrist and hand gestures was systematically investigated in two authentication modes: verification and identification. The results from 24 participants showed that the TD features significantly ( ${p} ) and consistently outperformed FDT and AR features for all channel numbers. The results also showed that the performance of a four-channel setup was not significantly different from those with higher number of channels. The average equal error rate (EER) for a four-channel sEMG verification system was 4% for TD features, 5.3% for FDT features, and 10% for AR features. For an identification system, the average Rank-1 error (R1E) for a four-channel configuration was 3% for TD features, 12.4% for FDT features, and 36.3% for AR features. The electrode position on the flexor carpi ulnaris (FCU) muscle had a critical contribution to the authentication performance. Thus, the combination of the TD feature set and a four-channel sEMG system with one of the electrodes positioned on the FCU are recommended for optimal authentication performance.

Volatility proxies like Realised Volatility (RV) are extensively used to assess the forecasts of squared financial return produced by Autoregressive Conditional Heteroscedasticity (ARCH) models. But are volatility proxies identified as expectations of the squared return? If not, then the results of these comparisons can be misleading, even if the proxy is unbiased. Here, a tripartite distinction between strong, semi-strong and weak identification of a volatility proxy as an expectation of squared return is introduced. The definition implies that semi-strong and weak identification can be studied and corrected for via a multiplicative transformation. Well-known tests can be used to check for identification and bias, and Monte Carlo simulations show they are well-sized and powerful -- even in fairly small samples. As an illustration, twelve volatility proxies used in three seminal studies are revisited. Half of the proxies do not satisfy either semi-strong or weak identification, but their corrected transformations do. Correcting for identification does not always reduce the bias of the proxy, so there is a tradeoff between the choice of correction and the resulting bias.

The application of the Bootstrap-Metropolis-Hastings algorithm is limited to fixed dimension models. In various fields, data often has a variable dimension model. The Laplacian autoregressive (AR) model includes a variable dimension model so that the Bootstrap-Metropolis-Hasting algorithm cannot be applied. This article aims to develop a Bootstrap reversible jump Markov Chain Monte Carlo (MCMC) algorithm to estimate the Laplacian AR model. The parameters of the Laplacian AR model were estimated using a Bayesian approach. The posterior distribution has a complex structure so that the Bayesian estimator cannot be calculated analytically. The Bootstrap-reversible jump MCMC algorithm was applied to calculate the Bayes estimator. This study provides a procedure for estimating the parameters of the Laplacian AR model. Algorithm performance was tested using simulation studies. Furthermore, the algorithm is applied to the finance sector to predict stock price on the stock market. In general, this study can be useful for decision makers in predicting future events. The novelty of this study is the triangulation between the bootstrap algorithm and the reversible jump MCMC algorithm. The Bootstrap-reversible jump MCMC algorithm is useful especially when the data is large and the data has a variable dimension model. The study can be extended to the Laplacian Autoregressive Moving Average (ARMA) model.

Identification of spatial outlier is essential in revealing hidden useful knowledge in different fields of statistical applications for big data. The Score Statistic (SCi) has been used as a diagnostic tool for the identification of spatial outliers in big data. Nonetheless, the SCi method suffers from masking and swamping effects. In order to reduce the swamping effect, we propose two methods denoted as tMDR and tEW. The tMDR and tEW methods adopt location adjacency to construct spatial weights, namely metric distance reciprocal (MDR) weight and exponential weight (EW), respectively, to detect outliers in spatial autoregressiveregressive model (SAR), spatial autoregressive error model (SEM) and general spatial autoregressive-regressive model (GSM). Difference between spatial residuals are calibrated to incorporate adjacency effect into spatial outlier residual. The results of the simulation study and real example show that the performances of the three methods are equally good for SAR model. The tMDR and tEW are comparable and both outperform the SCi for SEM and GSM models with less swamping effects and less computational running times.

Soft sensors have been extensively developed to estimate the difficult-to-measure quality variables for real-time process monitoring and control. Process nonlinearities and dynamics are two main challenges for accurate soft sensor modeling. To cope with these problems, two temporal convolutional network (TCN)-based soft sensor models are developed in this paper. With a hierarchy of temporal convolution kernels and large receptive fields, TCN is able to describe the nonlinearities and long dynamic dependence of process variables. Thus, a TCN model is first designed for nonlinear dynamic soft sensing. However, the original TCN model neglects the auto-correlations of the quality variable, as well as the cross-correlations between the quality variable and process variables. Hence, an autoregressive TCN (AR-TCN) is further constructed by integrating the lagged quality data in the past neighboring samples for quality prediction of the current sample. Thus, AR-TCN is able to capture the auto-correlations and cross-correlations between the quality variables and process variables, which are more beneficial for quality prediction. The effectiveness of two proposed models is validated on an industrial debutanizer column and an industrial hydrocracking process.

Microelectrode recording (MER) or microelectrode signals recording of local field potentials by means of subthalamic-nuclei deep brain stimulation is highly successful for construing or deducing Parkinson disease (PD) signal analysis acquiescent to elucidation are fetching ever more germane. These signals are supposed to emulate STN neurons action-potential movement and, these potential frequency modulations are coupled to spiking-events. The method uses auto regression stochastic (random nature) model machine learning approach and, other standard techniques as of system identification field. A usual conventional local field potential implication involves computing spectral-densities, i.e., power (P S Ds) of these signals—waveforms, that confines power on different frequencies. But, P S D s is second-order statistics might not confine non trivial temporal-dependencies which subsist in unprocessed data. Hence, we suggest L F Ps technique which is valuable in support of relating or unfolding distinctive plus sole features of temporal-dependencies in L F P waveforms. This technique is derived as of auto-regression (A R) modeling originating as of the systems plus control theory in fastidious systems-identification. We have distinctively, built A R models of L F P movement activity plus inferred, and also verified statistically major differentiations in temporal-dependencies among damaged tissues of nuclei plus protective areas in Parkinson`s receiving innocuous microelectrodes via deep brain stimulator (D BS). Differentiations in spectral-densities of field-waveforms amid the two conglomerates were not statistically-significant. Keywords: Microelectrode recording (MER), Parkinson disease (PD).

The influx of intensive longitudinal data creates a pressing need for complex modeling tools that help enrich our understanding of how individuals change over time. Multilevel vector autoregressive (mlVAR) models allow for simultaneous evaluations of reciprocal linkages between dynamic processes and individual differences, and have gained increased recognition in recent years. High-dimensional and other complex variations of mlVAR models, though often computationally intractable in the frequentist framework, can be readily handled using Markov chain Monte Carlo techniques in a Bayesian framework. However, researchers in social science fields may be unfamiliar with ways to capitalize on recent developments in Bayesian software programs. In this paper, we provide step-by-step illustrations and comparisons of options to fit Bayesian mlVAR models using Stan, JAGS and Mplus, supplemented with a Monte Carlo simulation study. An empirical example is used to demonstrate the utility of mlVAR models in studying intra- and inter-individual variations in affective dynamics.

Purpose Bulk shipping mostly facilitates the smooth flow of raw materials around the globe. Regardless, forecasting a bulk shipbuilding orderbook is a seldom researched domain in the academic arena. This study aims to pioneer an econophysics approach coupled with an autoregressive data analysis technique for bulk shipbuilding order forecasting. Design/methodology/approach By offering an innovative forecasting method, this study provides a comprehensive but straightforward econophysics approach to forecast new shipbuilding order of bulk carrier. The model has been evaluated through autoregressive integrated moving average analysis, and the outcome indicates a relatively stable good fit. Findings The outcomes of the econophysics model indicate a relatively stable good fit. Although relevant maritime data and its quality need to be improved, the flexibility in refining the predictive variables ensure the robustness of this econophysics-based forecasting model. Originality/value By offering an innovative forecasting method, this study provides a comprehensive but straightforward econophysics approach to forecast new shipbuilding order of bulk carrier. The research result helps shipping investors make decision in a capital-intensive and uncertainty-prone environment.

Mobile Crowdsensing (MCS) is a powerful sensing paradigm, which provides sufficient social data for cognitive analytics in industrial sensing, and industrial manufacturing. Considering the sensing costs, sparse MCS, as a variant, only senses the data in a few subareas, and then infers the data of unsensed subareas by the spatio-temporal relationship of the sensed data. Existing works usually assume that the sensed data are linearly spatiotemporal dependent, which cannot work well in real-world nonlinear systems, and thus, result in low data inference accuracy. Moreover, in many cases, users not only require inferring the current data, but also have an interest in predicting the near future, which can provide more information for users’ decision making. Facing these problems, we propose a deep learning-enabled industrial sensing, and prediction scheme based on sparse MCS, which consists of two parts: matrix completion and future prediction. Our goal is to achieve high-precision prediction of future moments under the hypothesis of sparse historical data. To make full use of the sparse data for prediction, we first propose a deep matrix factorization method, which can retain the nonlinear temporal-spatial relationship, and perform high-precision matrix completion. In order to predict the subareas’ data in several future sensing cycles, we further propose a nonlinear autoregressive neural network, and a stacked denoising autoencoder to obtain the temporal–spatial correlation between the data from different cycles or subareas. According to the results gained by experiments on four real-world industrial sensing datasets consisting of six typical tasks, it can be seen that the method in this article improves the accuracy of prediction using sparse data.

We propose a methodology for synthetic generation of solar irradiance (shortwave flux) and air temperature time series using copula functions. The use of copulas for the simulation gives flexibility to represent the serial stochastic variability of the solar irradiation and the air temperature affecting the photo-voltaic (PV) panel energy output. Moreover, it allows to have more control on the desired properties of the time series, not only in the temporal and cross-dependencies, but also in the marginal distributions. We use mixtures of zero mass adjusted density distributions to assess the nature of solar irradiance, alongside vector generalized linear models for the bivariate time series time marginal distributions. We found that the copula autoregressive methodology used, including the zero mass characteristics of the solar irradiance time series, accurately models the stochastic phenomena. Experimental analysis with observed data substantiates the usage and convenience of the proposed methodology to model solar irradiance time series and solar energy across the northern hemisphere, southern hemisphere and equatorial zones. These results will improve the understanding of the fluctuating nature of solar irradiance and also help to understand the underlying stochastic process of photo-voltaic energy production.

This study investigates the relationship between the financial market and the real business cycle in the US from February 1987 to March 2016. Using different monthly time-series as proxies for the financial and macroeconomic cycles, we first specify the determinants and then build two indicators to measure the financial and real business cycles based on principal component analysis. We identify not only the main different cycles for each indicator but also measure the duration of the phases for each indicator. Second, we study the relationship between economic and financial indicators per cycle and per phase using a vector autoregressive model. Our findings show that the economic indicator is useful to forecast the financial, and that there exists a significant relationship between the financial and economic cycles that is actively stronger during the “expansion–growth” phase. This result enables investors and policymakers to better forecast the future dynamics of financial sector using the information provided by the analysis of the real business cycle.

This research explores the possibility of monitoring apoptosis and classifying clusters of apoptotic cells based on the changes in ultrasound backscatter signals from the tissues. The backscatter from normal and apoptotic cells, using a high frequency ultrasound instrument are modeled through an autoregressive (AR) modeling technique. The proper model order is calculated by tracking the error criteria in the reconstruction of the original signal. The AR model coefficients, which are assumed to contain the main statistical features of the signal, are passed as the input to linear and nonlinear machine classifiers (Fisher linear discriminant, conditional Gaussian classifier, Naive Bayes classifier and neural networks with nonlinear activation functions). In addition, an adaptive signal segmentation method (least squares lattice filter) is used to differentiate the data from layers of different cell types into stationary parts ready for modeling and classification.

Forward-looking forecasting of the inflation rate could help the central bank and other government departments to better use monetary policy to stabilize prices and prevent the impact of inflation on market entities, especially for low- and middle-income groups. It can also help financial institutions and investors better make investment decisions. In this sense, the forecast of inflation rate is of great significance. The existing literature mainly uses linear models such as autoregressive (AR) and vector autoregressive (VAR) models to predict the inflation rate. The nonlinear relationship between variables and the mining of historical data information are relatively lacking. Therefore, the prediction strategies and accuracy of the existing literature need to be improved. The predictive model designed in deep learning can fully mine the nonlinear relationship between variables and process complex long-term time series dynamic information, thereby making up for the deficiencies of existing research. Therefore, this paper employs the recurrent neural networks with gated recurrent unit (GRU-RNN) model to train and analyze the Consumer Price Index (CPI) indicators to obtain inflation-related prediction results. The experimental results on historical data show that the GRU-RNN model has good performance in predicting China’s inflation rate. In comparison, the performance of the proposed method is significantly better than some traditional models, showing its superior effectiveness.

Spatial analytic approaches are classic models in econometric literature, but relatively new in social sciences. Spatial analysis models are synonymous with social network auto-regressive models which are also gaining popularity in the behavioral sciences. These models have two major benefits. First, dependent data, either socially or spatially, must be accounted for to acquire unbiased results. Second, analysis of the dependence provides rich additional information such as spillover effects. Structural Equation Models (SEM) are widely used in psychological research for measuring and testing multi-faceted constructs. So far, SEM that allow for spatial or social dependency are limited with regard to their flexibility, for example, when estimating nonlinear effects. Here, we provide a cohesive framework which can simultaneously estimate latent interaction/polynomial effects and account for spatial effects with both exogenous and endogenous latent variables, the Bayesian Spatial Auto-Regressive Structural Equation Model (BARDSEM). First, we briefly outline classic auto-regressive models. Next, we present the BARDSEM and introduce simulation results to exemplify its performance. Finally, we provide an empirical example using the spatially dependent extended US southern homicide data to show the rich interpretations that are possible using the BARDSEM. Finally, we discuss implications, limitations, and future research.

This study analyses the effect of natural gas consumption on economic performance in selected sub‐Saharan African countries between 1998 and 2017. The lag augmented vector autoregressive approach of Toda and Yamamoto to heterogeneous mixed panels illustrated in Emirmahmutoglu and Kose and the dynamic heterogeneous techniques are employed in data analysis. The study finds a significant long‐run relationship between the real gross domestic product and natural gas consumption. The study also reveals that natural gas consumption has a significant positive influence on economic performance in the short run. Furthermore, the results support the conservative hypothesis for the selected sub‐Saharan African countries. Therefore, it is reasonable for the selected sub‐Saharan African countries to formulate and promote cost‐effective and eco‐friendly energy and environmental policies that will enhance the use of natural gas without reducing economic performance in the long run.

This study aims to analyze the effect of the Central Bank, which is in Indonesia presented by Bank Indonesia (BI) Rate and inflation rate on the savings rate at commercial banks in Central Sulawesi by using the analysis tools of autoregressive conditional heteroscedasticity (ARCH) and generalized autoregressive conditional heteroscedasticity (GARCH). The results of the analysis show that the BI Rate and inflation have significant effects on saving rate at commercial banks in Central Sulawesi. The importance of being a source of investment funds to support development activities can be a consideration for the government to formulate a policy for allocating investment from these savings funds for productive sectors. Practically, it is hoped that the distribution of savings through investment activities can encourage economic growth and increase economic development in Central Sulawesi.

Many natural systems, especially biological ones, exhibit complex multivariate nonlinear dynamical behaviors that can be hard to capture by linear autoregressive models. On the other hand, generic nonlinear models such as deep recurrent neural networks often require large amounts of training data, not always available in domains such as brain imaging; also, they often lack interpretability. Domain knowledge about the types of dynamics typically observed in such systems, such as a certain type of dynamical systems models, could complement purely data-driven techniques by providing a good prior. In this work, we consider a class of ordinary differential equation (ODE) models known as van der Pol (VDP) oscil lators and evaluate their ability to capture a low-dimensional representation of neural activity measured by different brain imaging modalities, such as calcium imaging (CaI) and fMRI, in different living organisms: larval zebrafish, rat, and human. We develop a novel and efficient approach to the nontrivial problem of parameters estimation for a network of coupled dynamical systems from multivariate data and demonstrate that the resulting VDP models are both accurate and interpretable, as VDP's coupling matrix reveals anatomically meaningful excitatory and inhibitory interactions across different brain subsystems. VDP outperforms linear autoregressive models (VAR) in terms of both the data fit accuracy and the quality of insight provided by the coupling matrices and often tends to generalize better to unseen data when predicting future brain activity, being comparable to and sometimes better than the recurrent neural networks (LSTMs). Finally, we demonstrate that our (generative) VDP model can also serve as a data-augmentation tool leading to marked improvements in predictive accuracy of recurrent neural networks. Thus, our work contributes to both basic and applied dimensions of neuroimaging: gaining scientific insights and improving brain-based predictive models, an area of potentially high practical importance in clinical diagnosis and neurotechnology.

Accurate estimation of air transport demand is vital for airlines, related aviation companies, and government agencies. For example, both short-term and long-term business plans of airlines require accurate forecasting of future air traffic flows. This study aims to forecast the volume of air passengers in Kuwait International Airport (KIA), which is in the state of Kuwait. Using monthly air traffic volume data between January 2012 and December 2018, this study focuses on the modelling and forecasting the number of air passengers in KIA. A wide range of time series forecasting models are considered in this research, including autoregressive-integrated-moving average model (ARIMA), exponential smoothing with errors term (ETS), Holt-Winters exponential smoothing, neural network autoregression (NNAR), hybrid and Bayesian structural time series (BSTS), and a hybrid model. The forecasting performance of these models are compared using multiple train-test splits where the models are fitted on the training sets and evaluated on the test sets. The mean absolute percentage error (MAPE) is used to compare the performance of various models. Empirical analysis suggests that the BSTS model compares favorably against the other time series models in its ability to forecast complex time series. The BSTS model may be applied to study other complex time series forecasting problems with irregularity.

Air pollution can be attributed to the reduction in visibility, less agricultural activity, more health issues, and long-term destruction to infrastructure. This paper aimed to examine the validity of association among the Particulate matter (PM2.5) and number of acute upper respiratory infection (ARI) and Asthma (AS) patients using an autoregressive distributed lag (ARDL) approach. This ARDL model study was conducted in Lahore, Pakistan. We used monthly data of ARI and AS patients acquired from Directorate General Health Services Punjab and PM2.5 from Air Visual-IQAir during the period January 2018-August 2019. ARDL bound testing technique was used to investigate the association between number of AS, ARI patients and PM2.5. In the short run, the PM2.5 has substantial positive impact on number of AS patients in Lahore. The values of short-run coefficient depicts that the association between PM2.5 and ARI patients is stronger than AS. The effect of PM2.5 on number of patients in short term is more than that in the long-term. For both AS and ARI, in the long run, PM2.5 has negative impact on number of patients.

The article considers the problem of forecasting nonlinear nonstationary processes, presented in the form of time series, which can describe the dynamics of processes in both technical and economic systems. The general technique of analysis of such data and construction of corresponding mathematical models based on autoregressive models and recurrent neural networks is described in detail. The technique is applied on practical examples while performing the comparative analysis of models of forecasting of quantity of channels of service of cellular subscribers for a given station and revealing advantages and disadvantages of each method. The need to improve the existing methodology and develop a new approach is formulated.

Sequential variational autoencoders (VAEs) with a global latent variable z have been studied for disentangling the global features of data, which is useful for several downstream tasks. To further assist the sequential VAEs in obtaining meaningful z, existing approaches introduce a regularization term that maximizes the mutual information (MI) between the observation and z. However, by analyzing the sequential VAEs from the information-theoretic perspective, we claim that simply maximizing the MI encourages the latent variable to have redundant information, thereby preventing the disentanglement of global features. Based on this analysis, we derive a novel regularization method that makes z informative while encouraging disentanglement. Specifically, the proposed method removes redundant information by minimizing the MI between z and the local features by using adversarial training. In the experiments, we trained two sequential VAEs, state-space and autoregressive model variants, using speech and image datasets. The results indicate that the proposed method improves the performance of downstream classification and data generation tasks, thereby supporting our information-theoretic perspective for the learning of global features.

Transmission of natural gas from its sources to end users in various geographical locations is carried out mostly by natural gas transmission pipeline networks (NGTNs). Effective design and operation of NGTNs requires insights into their steady-state and, particularly, dynamic behavior. This, in turn, calls for efficient computer-aided approaches furnished with accurate mathematical models. The conventional mathematical methods for the dynamic simulation of NGTNs are computationally intensive. In this paper, the use of autoregressive neural networks for cost-effective dynamic simulation of NGTNs is proposed. Considering the length, diameter, roughness, and elevation as the main characteristics of a single pipeline, a neural network pipeline (NNPL) is designed and trained based on the data from a dynamic process simulator. Arbitrary NGTNs can then be easily constructed by connecting the developed NNPLs as the building blocks. The performance of the NNPL network is demonstrated through a number of benchmark pipeline systems, where a very good agreement with the benchmark results is found.

This study utilizes a time-varying parameter Bayesian vector autoregressive model to investigate the dynamic interactions between geopolitical risk (GPR) and renewable energy consumption growth (RECG). The identification strategy is flexible to accommodate cases both with and without restrictions of the direction of impact. It is shown that GPR shocks have positive impacts on RECG over time. In contrast, RECG shocks decrease GPR in the whole sample period. These results show that renewable energy is a useful tool to reduce geopolitical risks. Meanwhile, the increasing geopolitical risks tend to augment renewable energy consumption. We also provide the responses at different time horizons and during particular geopolitical events. The estimating results are robust when industrial production growth is used as a control variable. Lastly, several implications for economic policy making are discussed.

We prove the existence of limiting distributions for a large class of Markov chains on a general state space in a random environment. We assume suitable versions of the standard drift and minorization conditions. In particular, the system dynamics should be contractive on the average with respect to the Lyapunov function and large enough small sets should exist with large enough minorization constants. We also establish that a law of large numbers holds for bounded functionals of the process. Applications to queuing systems, to machine learning algorithms and to autoregressive processes are presented.

Wind farms have a focus role in environmentally friendly energy production. There are short-term estimates of wind speed in planning energy production in wind power plants. In this article, we analyzed the wind speed in the Istanbul Avcilar region by an artificial neural network method (ANN) and regression method. One of the methods commonly used in estimation processes is Nonlinear Autoregressive Exogenous (NARX). We divide the data into three parts 70%, 15%, and 15%, respectively, for learning, validation, and testing. We used the Levenberg-Marquardt (LM) algorithm for data network training. We compared the predicted wind speed with the measured and tested values. We used MATLAB software in the analysis of the model. We obtained system outputs and regression models of wind speed with artificial neural network simulations. Besides, we calculated the effect sizes. Scientific Research Grant of Istanbul University-Cerrahpasa [2071] This research was support by the Scientific Research Grant of Istanbul University-Cerrahpasa, [Grant number: 2071]

Understanding the interaction of debris flow with different arrangements of buildings and providing an accurate and real-time predictions for the corresponding flow distribution has become a challenging task. Two parts of this problem, namely new constructions and rapid reaction, have made this as a daunting topic. Although the numerical methods can provide precise predictions, their applicability due to the above-mentioned issues is challenging as one has to redo the computations when new buildings are added. In this paper, thus, a deep learning model is developed to efficiently predict the spatiotemporal distribution of debris flow as it interacts with different building arrangements. As such, we employed U-net as the base model, which is combined with the residual network to alleviate the gradient vanishing issue. An autoregressive strategy is proposed to help the model capture the dynamics of the debris flow, where the output at the previous timestep is used as the input to predict the current timestep. A synthetic dataset based on the numerical methods is used to train the residual U-net model. The results indicate that the proposed method can predict the general trend at different timesteps, and its robustness is verified using an extensive sensitivity analysis.

Autoregressive backprojection (ARBP) approximates the procedure of forming synthetic‐aperture radar (SAR) sequential images as an autoregressive (AR) process, avoiding redundant computation between adjacent images. Azimuth apodization for ARBP entails designing the aforementioned AR process, including AR order selection and AR model parameter estimation. However, the conventional azimuth apodization method for ARBP provides only empirical means to conduct such a procedure. Moreover, it cannot address the ringing effect from the infinite impulse response of the AR model, and consequently, other unwanted azimuth pulses will degrade the SAR images. This study discusses azimuth apodization for ARBP and proposes a scheme that includes methods for AR model order selection and parameter estimation. Firstly, matrix rank‐k approximation is adopted for AR model order selection. Then, a revised azimuth apodization procedure for parameter estimation is discussed. It introduces a penalty factor to render the ringing effect and model estimation error controllable. Moreover, experiments based on an open data set validate the proposed ARBP azimuth apodization scheme. In this way, the proposed azimuth apodization obtains a balance between computational complexity and imaging quality for ARBP.

In this article, we address the problem of inferring direct influences in social networks from partial samples of a class of opinion dynamics. The interest is motivated by the study of several complex systems arising in social sciences, where a population of agents interacts according to a communication graph. These dynamics over networks often exhibit an oscillatory behavior, given the stochastic effects or the random nature of the local interactions process. Inspired by recent results on estimation of vector autoregressive processes, we propose a method to estimate the social network topology and the strength of the interconnections starting from partial observations of the interactions, when the whole sample path cannot be observed due to limitations of the observation process. Besides the design of the method, our main contributions include a rigorous proof of the convergence of the proposed estimators and the evaluation of the performance in terms of complexity and number of sample. Extensive simulations on randomly generated networks show the effectiveness of the proposed technique.

PurposeIt is quite possible that financial institutions including life insurance companies would encounter turbulent situations such as the COVID-19 pandemic before policies mature. Constructing models that can generate scenarios for major assets to cover abrupt changes in financial markets is thus essential for the financial institution's risk management.Design/methodology/approachThe key issues in such modeling include how to manage the large number of risk factors involved, how to model the dynamics of chosen or derived factors and how to incorporate relations among these factors. The authors propose the orthogonal ARMA–GARCH (autoregressive moving-average–generalized autoregressive conditional heteroskedasticity) approach to tackle these issues. The constructed economic scenario generation (ESG) models pass the backtests covering the period from the beginning of 2018 to the end of May 2020, which includes the turbulent situations caused by COVID-19.FindingsThe backtesting covering the turbulent period of COVID-19, along with fan charts and comparisons on simulated and historical statistics, validates our approach.Originality/valueThis paper is the first one that attempts to generate complex long-term economic scenarios for a large-scale portfolio from its large dimensional covariance matrix estimated by the orthogonal ARMA–GARCH model.

Several activities regarding water resources management are dependent on accurate monthly streamflow forecasting, such as flood control, reservoir operation, water supply planning, hydropower generation, energy matrix planning, among others. Most of the literature is focused on propose, compare, and evaluate the forecasting models. However, the decision on forecasting approaches plays a significant role in such models’ performance. In this paper, we are focused on investigating and confront the following forecasting approaches: i) use of a single model for the whole series (annual approach) versus using 12 models, each one responsible for predicting each month (monthly approach); ii) for multistep forecasting, the use of direct and recursive methods. The forecasting models addressed are the linear Autoregressive (AR) and Periodic Autoregressive (PAR) models, from the Box & Jenkins family, and the Extreme Learning Machines (ELM), an artificial neural network architecture. The computational analysis involves 20 time series associated with hydroelectric plants indicated that the monthly approach with the direct multistep method achieved the best overall performances, except for the cases in which the coefficient of variation is higher than two. In this case, the recursive approach stood out. Also, the ELM overcame the linear models in most cases.

Актуальной проблемой развития и повышения конкурентоспособности нефтегазовой отрасли является решение задачи надежного и эффективного электроснабжения технологических объектов новых месторождений и магистральных нефтегазопроводов, территориально удаленных от центральной электрической сети. Одним из эффективных способов решения данной проблемы является применение ветроэнергетических установок. Первоочередной задачей проектирования и технико-экономического обоснования применения ветроэнергетических установок является прогноз изменения скоростей ветра в месте установки электростанции. Стохастическая природа ветра и его большая изменчивость во времени и пространстве определяют высокую сложность данной задачи, для решения которой используют методы математического моделирования. Известные модели скорости ветра на основе цепей Маркова, авторегрессии и скользящего среднего не позволяют осуществлять варьирование шага моделирования, что ограничивает возможность их применения для имитационного моделирования режимов ветроэнергетических установок и энергетических систем на их основе. В статье предложена модель скорости ветра на основе стохастического дифференциального уравнения, устраняющая данный недостаток. Цель: разработка модели скорости ветра на основе стохастического дифференциального уравнения дробного процесса Орнштейна-Уленбека c периодической функцией среднего значения, обеспечивающей моделирование статических и динамических режимов работы ветроэнергетической установки на различных временных интервалах. Методы: математическое и компьютерное моделирование с использованием программной среды MatLab/Simulink. Результаты. Предложенная модель, в отличие от известных моделей подобного типа, воспроизводит как суточные, так и сезонные вариации скорости ветра, а также долговременную корреляционную зависимость моделируемого процесса, что позволяет осуществлять моделирование рабочих режимов ветроэнергетических установок на различных временных интервалах с требуемой дискретизацией. Валидация модели выполнена с использованием данных климатических наблюдений скорости ветра из электронного архива Всероссийского института гидрометеорологической информации. Адекватность модели подтверждена хорошей степенью соответствия моделируемых траекторий скорости ветра с данными фактических наблюдений, зарегистрированных на 518 метеостанциях, расположенных на территории России. Relevance. The urgent task of the technical and economic development of the Northern and Eastern regions of Russia is to provide reliable and efficient power supply to consumers, geographically located in remote, hard-to-reach areas. The use of wind power plants is a promising way to solve this problem. The primary task of the design and feasibility study of the use of wind power plants is to predict changes in wind speeds at the site of the power plant installation. The stochastic nature of the wind and its spatiotemporal variability explain the high complexity of this problem, for the solution of which the methods of mathematical modeling are used. The known models of wind speed based on Markov chains, autoregressive moving-average and other discrete models do not allow varying the time step, which does not allow their use for simulation of operating modes of wind turbines and wind energy systems. The article proposes a model of wind speed based on the stochastic differential equation which eliminates this drawback. The aim of the study is to construct wind speed model based on fractional Ornstein-Uhlenbeck process with a periodic long-term mean, which provides modeling of static and dynamic modes of operation of a wind power plant at different time intervals. Methods: mathematical and computer modeling using Matlab/Simulink software environment. Results. The proposed model, in contrast to the known SDE-based models, is able to produce autocorrelated wind speed trajectories with long-term dependence, daily and monthly variations to perform more detailed simulation of the operating modes of wind turbines at different time intervals with the required time step. The model was tested using the data of climatic observations of wind speed obtained from electronic archive of the All-Russian Institute of Hydrometeorological Information. The model adequacy was evaluated by comparing characteristics of simulated wind speed trajectories with actual observations collected at 518 weather stations located on territory of Russia.

We develop a heterogeneous autoregressive model of downside volatility (HAR‐DV) model with structural changes (HAR‐DV‐SC) model to investigate the effects of structural changes on predicting downside volatility. Then we employ HAR‐DV and HAR‐DV‐SC models to forecast downside volatilities in S&P 500 index, crude oil, gold, copper, and soybean futures markets. The in‐sample analysis shows that structural changes contain in‐sample information for predicting downside volatility. The out‐of‐sample analysis indicates that the HAR‐DV‐SC model outperforms the HAR‐DV model, suggesting that structural changes contain incremental out‐of‐sample information on future downside volatility. These results are robust and have important implications for risk management of stakeholders.

Activity classification has been an interesting area of research for many years, to better understand human behavior. Recent advancements in embedded computing systems allowed the emergence of several state-of-art solutions for human activity classification using sensors of a smartphone. The sensors generate temporal sequences of observations for human activity, which is called as Multivariate Time Series (MTS). Current state-of-art solutions for human activity classification suffer from two major limitations: first, the length of testing MTS should be equal to the training MTS and second, the MTS should not have any faulty time series. In real-time applications, it is desirable to classify a human activity using an incomplete MTS as early as possible. In this work, we propose a fault-tolerant early classification of MTS (FECM) approach to address these limitations. FECM builds a set of classification models using MTS training dataset. The approach employs Gaussian Process classifier to estimate minimum required length of time series, which is used to predict a class label of new MTS. Further, FECM uses an Auto Regressive Integrated Moving Average model to identify faulty time series in the new MTS. Finally, we conduct an experiment to evaluate the performance of FECM using accuracy and earliness metrics.

Wind prediction technology has been the focus on the national research with the basis of the power system planning, a reference for power dispatch, and the optimal power flow distribution. Prediction technology is moving into the direction of controlling refinement with the development of the information technology, the artificial intelligence technology, and the improvement of edge computing devices. The wind farms can use real-time wind power prediction to improve their overall efficiency through advanced planning of the wind turbine adjustment and the pre-setting of the yaw, pitch, and generation excitation control systems. This paper proposes an innovative autoregressive dynamic adaptive (ARDA) model based on the improvement of the autoregressive (AR) model. The fixed parameter estimation method of the AR model is improved in the proposed model to a dynamically adaptive stepwise parameter estimation method. Meanwhile, the coefficients of the model are updated adaptively based on the characteristics of wind power data, which improves the accuracy of the proposed model. The prediction accuracy of the proposed model is further improved by the residual function. It was observed that the model adapts well to wind power data with different degrees of volatility. The ARDA model and two other models were tested by using stationary and fluctuating wind power data (unit: seconds), and the wind power prediction results at different forecasting step lengths were compared. It was observed that the ARDA model is more accurate, with faster calculation rate, and better dynamic adaptability to data fluctuations than the ARIMA and LSTM models. This paper proposes an important method for real-time power prediction that can be employed for the advanced control and improved the power generation of wind farms.

In this paper, we present a simple approach to train Generative Adversarial Networks (GANs) in order to avoid a \textit {mode collapse} issue. Implicit models such as GANs tend to generate better samples compared to explicit models that are trained on tractable data likelihood. However, GANs overlook the explicit data density characteristics which leads to undesirable quantitative evaluations and mode collapse. To bridge this gap, we propose a hybrid generative adversarial network (HGAN) for which we can enforce data density estimation via an autoregressive model and support both adversarial and likelihood framework in a joint training manner which diversify the estimated density in order to cover different modes. We propose to use an adversarial network to \textit {transfer knowledge} from an autoregressive model (teacher) to the generator (student) of a GAN model. A novel deep architecture within the GAN formulation is developed to adversarially distill the autoregressive model information in addition to simple GAN training approach. We conduct extensive experiments on real-world datasets (i.e., MNIST, CIFAR-10, STL-10) to demonstrate the effectiveness of the proposed HGAN under qualitative and quantitative evaluations. The experimental results show the superiority and competitiveness of our method compared to the baselines.