Correlated Time Series Research Articles

Forecasting air passenger traffic and market share using deep neural networks with multiple inputs and outputs.

In this study, we address the challenge of accurate time series forecasting of air passenger demand using historical market demand data from the U.S. commercial aviation industry in the 21st century. Commercial aviation is a major contributor to the U.S. economy, directly or indirectly generating ~US$1.37 trillion annually, or 5% of annual GDP, and supporting more than 10 million jobs (Airlines for America, 2024). Over 1 billion passengers flew through U.S. airports in 2023 (Bureau of Transportation Statistics, 2024a). Using multiple correlated time series inputs predicts future values of multiple interrelated time series and leverages their mutual dependencies to enhance accuracy. In this study, we introduce a two-stage algorithm employing a deep neural network for correlated time series forecasting, addressing scenarios where multiple input variables are interrelated. This approach is designed to capture the influence that one time series can exert on another, thereby enhancing prediction accuracy by leveraging these interdependencies. In the first stage, we fit four Recurrent Neural Network (RNN) models to generate accurate univariate forecasts, each functioning as a single input-output model to predict aggregated market demand. The Gated Recurrent Unit (GRU) model was the top performer for our dataset overall. In the second stage, we apply the best fitted model (GRU Model) from Stage 1 to each individual competitor (disaggregated from the market) and then merge all input tensors using the Concatenate function. We hope to contribute to the relevant body of knowledge with a deep neural network framework for forecasting market share among competitors in the U.S. commercial aviation industry, as no similar approach has been documented in the literature. Given the importance of the industry, there is potentially great value in applying sophisticated forecasting techniques to achieve accurate predictions of air passenger demand. Moreover, these techniques may have wider applications and can potentially be employed in other contexts.

Video anomaly detection with long-and-short-term time series correlations

Video anomaly detection with long-and-short-term time series correlations

Deciphering changes in the incidence of hemorrhagic stroke and cerebral venous sinus thrombosis during the coronavirus disease 2019 pandemic: A nationwide time-series correlation study.

Hemorrhagic stroke and cerebral venous sinus thrombosis (CVST) are associated with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and vaccination. We aimed to investigate changes in the incidence of hemorrhagic stroke and CVST in South Korea before and during the coronavirus disease 2019 pandemic and the factors associated with these changes. We conducted a nationwide time-series study using population-based databases between 2007 and 2022. The real-world and forecasted incidences of acute non-traumatic subarachnoid hemorrhage (SAH), intracerebral hemorrhage (ICH), and CVST during the pandemic period (2020-2022) were estimated and compared with the pre-pandemic period (2007-2019). The prevalence of conventional risk factors was measured using time-series data. Finally, a time-series correlation analysis was performed to examine the temporal association between conventional risk factors, SARS-CoV-2 infection, and SARS-CoV-2 vaccination. The incidence of hemorrhagic stroke (SAH and ICH) was lower during the pandemic than during the pre-pandemic period. This observed decrease was associated with a reduction in the prevalence of conventional risk factors but not with SARS-CoV-2 infection or vaccination. The incidence of CVST was higher during the pandemic than during the pre-pandemic period, which may be temporally related to SARS-CoV-2 vaccination (Pearson correlation coefficient [r] = 0.349, P = 0.031). We report reassuring evidence of hemorrhagic stroke associated with SARS-CoV-2 infection and vaccination. However, awareness of CVST may be required for future vaccine rollouts and SARS-CoV-2 outbreaks.

ИЗМЕНЕНИЯ КЛИМАТА НА ТЕРРИТОРИИ РЕСПУБЛИКИ ТАТАРСТАН

The paper looks at the dynamics of air temperature changes in Kazan in the 19th-21st centuries against the background of climatic processes occurring in the Northern hemisphere. The study used proven statistical methods of time series processing, correlation and trend analysis, and a Potter filter to isolate the low-frequency component. The fastest rate of warming in the region was established to occur in the winter-spring period. At the same time, the three decades between 1991 and 2021 turned out to be warmer in each month than the 5 previous three-decade periods. The contribution of the entire Northern Hemisphere to the fluctuations in the air temperature in Kazan was estimated (42% in winter and 27% in summer). It has been established that the average annual temperatures in Kazan and the entire Northern Hemisphere changed in the opposite phase until 1970. Since 1870, the winter temperature in Kazan has increased by 4.7°C. During the summer period, a ‘hump’ (warming) was detected on the low–frequency temperature curve in Kazan in the period 1920–1960, which can be explained by the fact that during this period the atmosphere in the Northern Hemisphere was more transparent (by 2%) than in the other periods. The consequences of climate change in the region in recent decades have had an effect on the dynamics of agro-climatic resources of Tatarstan. Several agrometeorological indicators were calculated, including the M.I. Budyko dryness index, the Selyaninov hydrothermal coefficient (HTC), and the Sapozhnikova moisture coefficient. According to the estimates made, there has been an increase in the duration of the growing season and the sum of effective temperatures, and also a weak trend toward increasing aridity in the summer period has emerged. Using ensemble calculations based on 40 climate models from the CMIP6 project, the paper provides an assessment of future climate changes in the region: the annual temperature in Kazan is expected to increase by 2.4°C by the end of the 21st century according to the ssp245 scenario. Under the ‘toughest’ ssp585 scenario, this increase will be 5.5°C.

2001–2020年中国东北区域土壤水蚀数据集

The black soil area in Northeast China plays an important role in food production and ecological security in China. However, over-cultivation practices have led to severe soil erosion, which seriously threatens food security and ecological environment in Northeast China. Accurate simulations of water erosion with high spatio-temporal resolution are important for advancing sustainable development goals, such as promoting sustainable agriculture and monitoring land degradation in Northeast China. This dataset is based on Google Earth Engine (GEE) cloud platform, integrating multi-source remote sensing data, deconstructing RUSLE model, and optimizing algorithm combinations for each factor of the model. We compared and verified the annual sand transport volume and sand transport modulus data from the main hydrological monitoring stations in Northeast China with the soil erosion modulus estimated by the RUSLE model. Then, we selected the optimal factor algorithm combination based on three accuracy metrics (time series correlation, root mean square error and mean absolute error), and obtained the estimation results of soil water erosion modulus with the resolution of 250 m. This dataset can better depict the spatial distribution and temporal changes of soil erosion modulus in Northeast China from 2001 to 2020. It can serve as an effective reference for soil erosion control and assessments in Northeast China.

Research on Physical Education Management System in Higher Education Institutions in the Context of Deep Learning

Abstract This paper focuses on the management of physical movement sports disciplines and combines relevant algorithms to design an artificial intelligence-based physical education management system framework. In terms of related algorithms, the first is for the research of character recognition algorithms. Considering the intra-class differences between individual movements, the LR-based feature extraction algorithm is proposed. Secondly, for the research of the action evaluation algorithm, the statistical value, DTW parameter action time series difference and correlation coefficient based on the statistical value and DTW parameter are used as the features, combined with SVM-based feature extraction algorithm for perceiving the action differences. Finally, with the help of deep neural networks for movement analysis, we can improve students’ movement techniques during training. In the course schedule supported by the system, the proportion of moderate-intensity exercise time is high, basically, more than 40%, which meets the WHO requirements for high-intensity exercise time in physical education classes. The post-course RPE feedback from students was spread out between 12 and 15, and the highest post-course RPE value was 14.82<unk>1.38, which was evaluated within a reasonable range of intervals. The teaching management system in this paper can meet the requirements of physical training and teaching, and provide quantifiable basis and management tools for the informatization and precise teaching of physical movement subjects.

The Relationship Between Wildfires and Respiratory Health

Wildfires have been increasing in length and frequency since the mid-1980s, which emit pollutant matter that can adversely impact human health, specifically exacerbating negative respiratory impacts (Lipner et al., 2019). Although many studies examine the respiratory impacts of worsened wildfire conditions, a limited number take place outside of the western United States, occur over prolonged periods, and look at how air quality is changing due to warming climates. The objective of this study is to investigate how air quality health index (AQHI) values have changed over a period of two decades (January 1, 2001 to December 31, 2021) in the Great Plains Air Zone (GPAZ) in southern Saskatchewan and the potential respiratory implications of AQHI changes. The study uses a mixed methods approach: time series analysis, correlation analysis, and regression analysis to analyze AQHI and a literature review to study respiratory implications of AQHI increases. We found a significant increase in AQHI over time of 19.3% (p < 0.0001, n = 7308) and three main outcomes for respiratory data: (1) individuals with respiratory conditions have increased susceptibility to worsened air conditions, (2) elevated AQHI values cause increases in healthcare admittances for respiratory conditions such as asthma and COPD, and (3) women and individuals in the age ranges of 15-65 are particularly susceptible to respiratory outcomes from wildfire smoke. The study provides areas for future research, including the implementation of apps to track respiratory outcomes, the impact of other risk factors on respiratory health, and the effects of AQHI on Indigenous populations.

Tourism, Inequality and Segregation in The North Lombok Regency

The aim of this research is to look at the effects of tourism spatially and examine its impact on the aggregate welfare of the population. Therefore, the author asks two fundamental questions in this research. First, is the rapid demand for tourism capable of creating new regional growth poles spatially? Second, is increasing tourism receipts able to create welfare effects. The method used in this research is simple time series correlation and regression. The findings in this research are that the dependent variables tend to have a strong influence on education and employment indicators. Tourism seems to only have a false effect on economic activity in North Lombok Regency. It is proven that since this district was expanded, there have been no significant growth poles in this region. Regional and economic segregation is still clearly visible today.

Impact of Real-Time Data, Market Sentiment, and Economic Factors on Investment Profitability in Indonesia

The study investigates the complex interplay of real-time data, market sentiment, and economic factors in shaping investment profitability in the context of Indonesia. Real-time data, including stock prices and trading volumes, market sentiment derived from sentiment analysis and expert interviews, and economic factors such as GDP growth, inflation rates, and interest rates are examined for their impact on investment outcomes. Quantitative methods, including correlation analysis, regression analysis, and time series analysis, are employed to analyze a comprehensive dataset spanning an extended period. The findings reveal that real-time data has a significant positive influence on investment profitability, underlining the importance of timely access to financial information. Market sentiment also exerts a notable impact, as more positive sentiment correlates with higher profitability. However, the relationship between economic factors and profitability is nuanced, with GDP growth exhibiting a positive influence and inflation rates and interest rates demonstrating weak negative effects. These results offer practical insights for investors and financial analysts, emphasizing the need to consider a holistic approach that incorporates real-time data, sentiment analysis, and macroeconomic indicators in investment decision-making. The study contributes to a deeper understanding of investment dynamics in the emerging market of Indonesia.

Correlation analysis of heart rate variations and glucose fluctuations during sleep

ObjectiveThe body's glucose concentration is influenced by carbohydrate intake, insulin-induced carbohydrate reduction, and hepatic glycogen breakdown induced by stress hormones. This study investigated the potential of employing glucose fluctuations as a measure of stress by examining the relationship between heart rate variability (HRV) data and glucose levels during sleep in healthy subjects. MethodsIn this cross-sectional study, a chest-worn electrocardiogram (ECG) and continuous glucose monitoring device (CGM) were respectively used to monitor the heart rate intervals and glucose fluctuations of five subjects (two males, three females) during sleep. A time-series correlation analysis was performed on the HRV data extracted from heart rate intervals and the corresponding glucose fluctuation data. ResultsThe time-series analysis of ECG and CGM data collected from subjects during sleep (n = 25 nights) revealed a moderate negative correlation between glucose levels and HRV, with a cross-correlation coefficient of r = −0.453. ConclusionSimilar to HRV, changes in stress levels can be detected by observing glucose fluctuations, particularly during sleep when the impact of food intake can be eliminated. Our findings highlight a significant correlation between glucose levels and HRV, indicating that glucose fluctuations can be used as an indicator of autonomic nervous system activity in an exploratory study.

Multi-attention Generative Adversarial Network for multi-step vegetation indices forecasting using multivariate time series

Generative Adversarial Networks (GANs) are one of the most significant research directions in the field of Deep Learning (DL). GANs has received wide attention due to their outstanding ability to produce realistic-looking images and can implicitly learn rich distributions over Spatio-Temporal (ST) correlations in Multivariate Time Series (MTS). Forecasting using MTS remains a challenge that needs to be solved to make successful tasks in different ST related applications such as vegetation forecasting. Vegetation is an extremely valuable component of our global ecosystem, as well as an important indicator of land cover change dynamics and productivity. As a result, the Normalized Difference Vegetation Index (NDVI) is one of the most widely used indices in vegetation related studies. The challenges of accurately forecasting vegetation indices come from both the non-stationary ST correlations of the NDVI data and the complex impact of the external context factors, including weather and remote sensing indices. It is further more challenging for most existing one-step forecasting methods, making an accurate forecast over many future time slots. To achieve this goal, we propose a new multi-attention GAN model composed of three main parts. An encoder network first encodes the driving sequence into latent space vectors and store all the information carried by the driving sequence. A generator is then in charge of generating forecasting data by extracting long-term temporal patterns. Eventually, in order to increase forecasting accuracy, we use as a third part an improved discriminator for classification and feedback purposes. The suggested model has been extensively tested with real-world datasets, demonstrating its effectiveness and robustness compared to existing methods. The test results indicate high performance with a Coefficient of Determination (R2) of 0.95, Root Mean Square Error (RMSE) of 0.04, Mean Absolute Error (MAE) of 0.01, and Mean Absolute Percentage Error (MAPE) of 15.35.

Sea clutter radar target detector based on autoregressive sieve bootstrap

The autoregressive sieve (ARS) bootstrap method can capture the correlation of time series and expand the original data set to the required volume. In this study, we introduce the ARS bootstrap into the matrix constant false alarm rate (CFAR) method to counter the inability of the Monte Carlo methods to extract the detection threshold due to the limited number of pulses under a certain false alarm probability. We assume the echo data of each range cell in a coherent processing interval (CPI) to be a stationary time series, and regard target detection as a two-sample hypothesis testing problem. We propose a statistic to test the equivalence of the autocovariance corresponding to two potential time series by using the correlation between pulses in a short pulse dataset, and analyze the validity of the ARS bootstrap for this statistic. The statistic distribution is obtained according to the bootstrap procedure and the threshold is obtained under fixed false alarm probabilities. Numerical experiments on the simulated and real sea clutter data were conducted to compare the detection results of the partial matrix CFAR detectors for the short and long pulse cases. With the introduction of the bootstrap method, we do not need to model the clutter amplitude distribution and the related statistical characteristics during target detection. To improve the detection probability with low signal-to-clutter ratio (SCR), we propose the two-sample autocovariance vector (ACV) detector based on the proposed statistic and use the correlation properties between pulses. We observed a 28% increase in detection probability when the SCR was 0 dB compared to other matrix CFAR methods, such as detectors based on Logarithmic Euclidean (LE) distance, Kernel Function-based Symmetric Kullback-Leibler (KSKL) divergence, and Total Bregman divergence (TBD) for the short pulse case.

Power system data-driven dispatch using improved scenario generation considering time-series correlations

Reinforcement learning (RL) is recently studied for realizing fast and adaptive power system dispatch under the increasing penetration of renewable energy. RL has the limitation of relying on samples for agent training, and the application in power systems often faces the difficulty of insufficient scenario samples. So, scenario generation is of great importance for the application of RL. However, most of the existing scenario generation methods cannot handle time-series correlation, especially the correlation over long time scales, when generating the scenario. To address this issue, this paper proposes an RL-based dispatch method which can generate power system operational scenarios with time-series correlation for the agent’s training. First, a time-generative adversarial network (GAN)-based scenario generation model is constructed, which generates system operational scenarios with long- and short-time scale time-series correlations. Next, the “N-1” security is ensured by simulating “N-1” branch contingencies in the agent’s training. Finally, the model is trained in parallel in an actual power system environment, and its effectiveness is verified by comparisons against benchmark methods.

A multiattribute financial time series forecast model based on double hierarchy fuzzy linguistic term set

Based on the double hierarchy linguistic term sets (DHLTS), a novel forecasting model is proposed considering both the internal fluctuation rules and the external correlation of different time series. The innovative aspects of this model consist of: (i) It can expresses more internal fluctuation and external correlation information, providing guarantees for improving the predictive performance of the model. (ii) The equivalent transformation function of DHLTS reduces the fuzzy granularity and improves the prediction accuracy. (iii) The application of similarity measures can extract the closest rules from historical states based on the distance operators of DHLTS. In addition, experiments on TAIEX considering the impact of the U.S. stock market and other data show that the model has good predictive performance.

A Spatiotemporal Fusion Model of Land Surface Temperature Based on Pixel Long Time-Series Regression: Expanding Inputs for Efficient Generation of Robust Fused Results

Spatiotemporal fusion technology effectively improves the spatial and temporal resolution of remote sensing data by fusing data from different sources. Based on the strong time-series correlation of pixels at different scales (average Pearson correlation coefficients > 0.95), a new long time-series spatiotemporal fusion model (LOTSFM) is proposed for land surface temperature data. The model is distinguished by the following attributes: it employs an extended input framework to sidestep selection biases and enhance result stability while also integrating Julian Day for estimating sensor difference term variations at each pixel location. From 2013 to 2022, 79 pairs of Landsat8/9 and MODIS images were collected as extended inputs. Multiple rounds of cross-validation were conducted in Beijing, Shanghai, and Guangzhou with an all-round performance assessment (APA), and the average root-mean-square error (RMSE) was 1.60 °C, 2.16 °C and 1.71 °C, respectively, which proved the regional versatility of LOTSFM. The validity of the sensor difference estimation based on Julian days was verified, and the RMSE accuracy significantly improved (p < 0.05). The accuracy and time consumption of five different fusion models were compared, which proved that LOTSFM has stable accuracy performance and a fast fusion process. Therefore, LOTSFM can provide higher spatiotemporal resolution (30 m) land surface temperature research data for the evolution of urban thermal environments and has great application potential in monitoring anthropogenic heat pollution and extreme thermal phenomena.

An alternative method for estimating Hurst exponent of control signals based on system dynamics

One of the indices for evaluating the performance of control systems based on the routine operation data is the Hurst index. It is shown that the Hurst index value is very close to that of the minimum variance index which measures the performance of a control loop in terms of minimum variance benchmark. The advantage of Hurst index over minimum variance index is that it does not require the system model and its parameters. The Hurst index is based on the Hurst exponent, which measures the time series correlations. This paper analytically scrutinizes the effect of parameters and dynamics of a stochastic control system on the output Hurst exponent, which helps better comprehend the Hurst index and its relationship with the minimum variance index. For this purpose, the control system output is modeled as an Autoregressive Moving Average (ARMA) process, and the relationship between the parameters of the ARMA process and the output Hurst exponent is examined. Furthermore, the characteristics of the autocorrelation function of a signal under minimum variance conditions are extracted. This paper also proposes a new technique for estimating the Hurst exponent of control signals based on the autocorrelation function. Simulation examples illustrate the efficiency of the proposed approaches.

Anomaly detection in telemetry data using a jointly optimal one-class support vector machine with dictionary learning

Anomaly detection based on telemetry data is a major issue in satellite health monitoring, given that it can identify unusual or unexpected events to avoid serious accidents and ensure the safety and reliability of operations. This study proposes a jointly optimal one-class support vector machine with dictionary learning (JODL-OC) framework for multivariate anomaly detection in telemetry. Dictionary learning (DL) extracts the correlation and local dynamics of time series into a sparse matrix, and the one-class support vector machine (OCSVM) detects anomalies in the transformed sparse space. The optimisation objective of DL is modified by introducing a novel regularisation term called soft-boundary OCSVM, so that the JODL-OC seamlessly integrates DL, sparse representation, and classifier training into a unified model to jointly learn the features and decision boundary. Furthermore, we propose an iterative approach to solve the optimisation function and obtain the optimal JODL-OC framework for anomaly detection. A case analysis based on satellite antenna telemetry data reveals that the proposed method improves the detection precision, F1-score, FPR and FAR compared to other existing anomaly detection methods, particularly to illustrate the superior performance in correlation anomalies.

NTformer: Near Time Transformer for multi-step prediction of wellhead pressure in fracturing operations

Fracturing process technology is a major factor in the effectiveness of fracturing for increasing production. One of the most critical issues during the construction of fracturing is the control of pressure variation, as excessive pressure can lead to serious accidents such as fracturing fluid leakage and well-bore control loss. To prevent accidents and improve operational efficiency, it is necessary to predict the future multi-step fracturing pressure in advance. Since fracturing operations are a process based on continuous changes in time dimensions, and their feature information has a high correlation in time series, this article optimizes a Transformer network with a serialization structure for wellhead pressure prediction, called Near Time Transformer (NTformer), which can effectively forecast pressure values across multiple time steps without relying on intricate feature engineering or domain-specific knowledge. Firstly, data cleaning and temporal encoding are performed on the original data to reduce the complexity of features at different stages and solve the problem of data temporal misalignment. To better learn from features, a feature conversion mechanism is proposed that the data features after embedding are divided into domain feature and wave feature. Further, utilizing the attention concept to spatially correlate feature information, and using the Transformer structure to transform features to achieve forecasting of future multi-step pressure. Finally, the idea of regularization is used to limit the parameters of the model to avoid over-fitting. Experimental results show that the method has a high prediction accuracy and can satisfy the accuracy requirements of the field application, offering a new ancillary method for improving the efficiency of oil and gas well fracturing construction and preventing safety accidents.

Feasibility and Utility of a Smartphone Application-Based Longitudinal Cough Monitoring in Chronic Cough Patients in a Real-World Setting.

This study evaluated the feasibility and utility of longitudinal cough frequency monitoring with the Hyfe Cough Tracker, a mobile application equipped with cough-counting artificial intelligence algorithms, in real-world patients with chronic cough. Patients with chronic cough (> 8-week duration) were monitored continuously for cough frequency with the Hyfe app for at least one week. Cough was also evaluated using the Leicester Cough Questionnaire (LCQ) and daily cough severity scoring (0-10). The study analyzed adherence rate, the correlation between objective cough frequency and subjective scores, day-to-day variability, and patient experience. Of 65 subjects consecutively recruited, 43 completed the study. The median cough monitoring duration was 13.9days, with a median adherence of 91%. Study completion was associated with baseline cough severity, and the adherence rate was higher in younger subjects. Cross-sectional correlation analyses showed modest correlations between objective and subjective cough measures at the group level. However, in time series correlation analyses, correlations between objective and subjective measures widely varied across individuals. Cough frequency had greater day-to-day variability than daily cough severity scores in most subjects. A patient experience survey found that 70% of participants found the cough monitoring helpful, 86% considered it acceptable, and 84% felt it was easy to use. Monitoring cough frequency longitudinally for at least one week may be feasible. The substantial day-to-day variability in objective cough frequency highlights the need for continuous monitoring. Grasping the implications of daily cough variability is crucial in both clinical practice and clinical trials.

Medium-term trends in economic and technological development of metals industry regions

Since metals industry regions are dependent on the location of production resources and infrastructure, as well as have highly capital-intensive production, they usually feature a slowly changing industrial and technological structure of production. In line with the hypothesis of the study, the technological structure of the manufacturing industry determines the differences in their economic development. The study aims to explore the medium-term trends and patterns in economic and technological development of the metals industry regions. The methodological basis of the research is the structural dynamic approach to studying regional economic development. The methods include multivariate data analysis and clustering algorithms, time series estimation and correlation analysis. The study uses the data on large and medium-sized enterprises by types of economic activities for 2006–2021 taken from an information and analytics system FIRA PRO. According to the findings, the Chelyabinsk, Sverdlovsk, Lipetsk, Vologda and Tula oblasts are metals industry regions. The identified medium-term trends in their economic development show different dynamics of input and output indicators of production both by all types of economic activities and in metals industry in particular. The study proves that the dynamics of metals industry regions’ development largely depends on technological characteristics of the manufacturing industry. The results of the research contribute to the understanding of the adaptation mechanisms of regions’ industry to the changing conditions of external environment.