Time Series Data Research Articles

Gaussian Process Phase Interpolation for estimating the asymptotic phase of a limit cycle oscillator from time series data

Gaussian Process Phase Interpolation for estimating the asymptotic phase of a limit cycle oscillator from time series data

Real-time mechanism-based interventions for daily alcohol challenges: Protocol for ecological momentary assessment and intervention.

Advancing evidence-based, tailored interventions for substance use disorders (SUDs) requires understanding temporal directionality while upholding ecological validity. Previous studies identified loneliness and craving as pivotal factors associated with alcohol consumption, yet the precise directionality of these relationships remains ambiguous. This study aims to establish a smartphone-based real-life intervention platform that integrates momentary assessment and intervention into everyday life. The platform will explore the temporal directionality of contextual influences on alcohol use among individuals experiencing loneliness and craving. We will target 180 individuals aged 18 to 70 in Germany who report loneliness, alcohol cravings, and meet risk or binge drinking criteria (over 14 standard drinks per week or five drinks in a single day for males, and over seven drinks per week or four drinks in a single day for females). Using a Within-Person-Encouragement-Design and Just-In-Time-Adaptive-Interventions, we will manipulate the contexts of loneliness and alcohol craving with cognitive reappraisal and physical activity interventions against a control condition (working memory task). Recruitment started in June 2024, with data collection and processing expected by June 2027. Our real-life intervention platform endeavors to serve as a robust tool for discerning the directionality of the effects from time series data in everyday life influences on alcohol use for the future study. Ultimately, it will pave the way for low-threshold prevention, clinical treatment, and therapy to target diverse contexts of everyday life in SUD. German Clinical Trials Register DRKS00033133.

Soil moisture retrieval over soybean fields during growing season using polarimetric decomposition and time-series RCM data

Soil moisture retrieval over soybean fields during growing season using polarimetric decomposition and time-series RCM data

Coastal vulnerability to extreme weather events: An integrated analysis of erosion, sediment movement, and land subsidence based on multi-temporal optical and SAR satellite data.

Coastal areas undergo continuous transformations, altering their geometry under the influence of external forces like tides, waves, and extreme events. Thus, monitoring the impact of extreme weather events on coastal regions is crucial to prevent potential cascading hazards. Here, we utilized time-series optical and SAR satellite data and tide records, coupled with sophisticated analytical techniques, to analyze erosion processes, sediment transport, and vertical land movement (VLM) at an embayed sandy beach (i.e., Haeundae Beach, South Korea). We found a significant erosion rate of -4.12m/year during typhoon periods, compared to non-typhoon periods (-3.61m/year). Between 2017 and 2023, 5.63ha of beach area and 14,428.9m3 of sediment were lost due to typhoon-induced sand movement, leading to coastline retreat and foreshore erosion. The PS-InSAR-derived localized VLM trends of fast-subsiding areas (-3.57mm/year), combined with a current sea-level rise (2.34mm/year), underscores the heightened vulnerability to future extreme weather events. The findings of this study provide insights into the erosional processes, land subsidence, and sediment transport due to extreme weather events, which can contribute to the development of erosion control technologies and engineering applications to prevent cascading hazards. The proposed integrated framework can be implemented in similar coastal environments to mitigate climate-change-induced coastal vulnerabilities.

FT-SMNet: Fourier transform sparse matrix network for structural health monitoring time series data forecasting

FT-SMNet: Fourier transform sparse matrix network for structural health monitoring time series data forecasting

PNC-based trend feature extraction method for time series data

PNC-based trend feature extraction method for time series data

Digital ecosystem for FAIR time series data management in environmental system science

Digital ecosystem for FAIR time series data management in environmental system science

Validation of a transdiagnostic psychopathology ecological momentary assessment protocol in a university student sample.

Ecological momentary assessment (EMA) collects real-time data in daily life, enhancing ecological validity and reducing recall bias. An EMA questionnaire that measures symptoms and transdiagnostic factors was recently developed with network modeling purposes. This study examines this EMA protocol's (a) subjective experience (e.g., burden, item clarity, survey frequency adequacy); (b) compliance, dropout, and predictors thereof; (c) the variability of EMA items across and within participants; and (d) the relations between EMA items and baseline standardized psychopathology questionnaires. University students (n = 262, Mage = 21.9, 84.8% females, 17.2% Dutch) completed eight daily momentary surveys (with the first including the morning survey), an evening survey, and a weekly survey during a 4-week EMA protocol. Additionally, a concluding survey examined participants' subjective experiences. Perceived burden was 3.40 on a 7-point scale, and people with higher levels of psychopathology found it more burdensome and more difficult to complete. Moreover, 67% of the surveys were completed, and 16% of the participants dropped out. Baseline psychopathology was not significantly associated with dropout or compliance. Moreover, surveys triggered in later study days, during the weekend, longer surveys, and surveys with lower financial reward were more likely to be missed. Between-subjects and within-subjects variability and correlations with baseline psychopathology varied across EMA items, with most EMA items showing sufficient within-individual variability for network modeling purposes and showing correlations across all types of psychopathology and transdiagnostic factors. The results suggest that the collection of intensive time-series data is feasible, and data quality and characteristics match requirements of different network models. (PsycInfo Database Record (c) 2025 APA, all rights reserved).

Reduced physiology-induced temporal instability achieved with variable-flip-angle fast low-angle excitation echo-planar technique with multishot echo planar time-resolved imaging.

Echo planar time-resolved imaging (EPTI) is a new imaging approach that addresses the limitations of EPI by providing high-resolution, distortion- and T2/ blurring-free imaging for functional MRI (fMRI). However, as in all multishot sequences, intershot phase variations induced by physiological processes can introduce temporal instabilities to the reconstructed time-series data. This study aims to reduce these instabilities in multishot EPTI. In conventional multishot EPTI, the time intervals between the shots comprising each slice can introduce intershot phase variations. Here, the fast low-angle excitation echo-planar technique (FLEET), in which all shots of each slice are acquired consecutively with minimal time delays, was combined with a variable flip angle (VFA) technique to improve intershot consistency and maximize signal. A recursive Shinnar-Le Roux RF pulse design algorithm was used to generate pulses for different shots to produce consistent slice profiles and signal intensities across shots. Blipped controlled aliasing in parallel imaging simultaneous multislice was also combined with the proposed VFA-FLEET EPTI to improve temporal resolution and increase spatial coverage. The temporal stability of VFA-FLEET EPTI was compared with conventional EPTI at 7 T. The results demonstrated that VFA-FLEET can provide spatial-specific increase of temporal stability. We performed high-resolution task-fMRI experiments at 7 T using VFA-FLEET EPTI, and reliable BOLD responses to a visual stimulus were detected. The intershot phase variations induced by physiological processes in multishot EPTI can manifest as specific spatial patterns of physiological noise enhancement and lead to reduced temporal stability. The VFA-FLEET technique can substantially reduce these physiology-induced instabilities in multishot EPTI acquisitions. The proposed method provides sufficient stability and sensitivity for high-resolution fMRI studies.

Decentralized Framework for Securing Smart Grids Using Blockchain and Machine Learning

Abstract: The transition to smart grids has revolutionized energy distribution, enabling more efficient and flexible power management through advanced communication and control systems. However, this interconnected structure makes smart grids vulnerable to cyberattacks, such as False Data Injection Attacks (FDIA), Distributed Denial of Service (DDoS) attacks, and data manipulation. These threats undermine the stability and reliability of the grid, and existing centralized security frameworks are often ill-equipped to address them due to their susceptibility to single points of failure and limited scalability. To overcome these challenges, this paper introduces a decentralized security framework that combines blockchain technology with machine learning (ML). The framework leverages blockchain to provide a transparent, immutable, and decentralized ledger, employing consensus mechanisms like Practical Byzantine Fault Tolerance (PBFT) or Proof of Authority (PoA) to ensure secure data validation. Alongside this, ML models, including Long Short-Term Memory (LSTM) networks and Convolutional Neural Networks (CNN), are used to detect anomalies in time-series data, such as FDIA, with high precision. Smart contracts embedded in the blockchain enable automated, real-time responses to threats, such as isolating compromised nodes or rerouting energy flows to maintain grid stability. Through simulations replicating real-world cyberattacks, the proposed framework demonstrated over 95% detection accuracy, a 30% reduction in response times, and enhanced computational efficiency with lower energy consumption. These results affirm the effectiveness of the framework as a scalable and resilient solution for modern smart grid security.

Forecasting the number of rural migrant workers in Chongqing using the optimized grey N_Verhulst model

Scientific prediction of migrant worker numbers provides decision-making references for resolving rural talent supply issues. Based on the evolutionary patterns and data features of Chongqing’s migrant workers, a new grey prediction model is constructed. The new model is constructed by introducing fractional-order operators in the real domain. In this way, the accumulating order of the traditional N_Verhulst model is optimized. It expands from 1 to all real numbers, thus enhancing its capacity to mine approximately saturated S-shaped time-series data. When the new N_Verhulst model is applied to simulate and predict migrant worker numbers, after optimizing the accumulating order, the mean relative simulation percentage error of the N_Verhulst model reduces from 3.66 to 2.93%, the mean relative forecasting percentage error from 8.02 to 2.18%, and the comprehensive mean relative percentage error from 4.53 to 2.78%. This shows that the optimization boosts the simulation and prediction performance of the N_Verhulst model. The prediction results show that the number of migrant workers in Chongqing will experience an orderly growth, rising from 2.41 million in 2023 to 2.85 million in 2028, with an increase of 18.26% and an average annual growth rate of 3.41%.

Uncovering temporal patterns in visualizations of high-dimensional data

With the increasing availability of high-dimensional data, analysts often rely on exploratory data analysis to understand complex data sets. A key approach to exploring such data is dimensionality reduction, which embeds high-dimensional data in two dimensions to enable visual exploration. However, popular embedding techniques, such as t-SNE and UMAP, typically assume that data points are independent. When this assumption is violated, as in time-series data, the resulting visualizations may fail to reveal important temporal patterns and trends. To address this, we propose a formal extension to existing dimensionality reduction methods that incorporates two temporal loss terms that explicitly highlight temporal progression in the embedded visualizations. Through a series of experiments on both synthetic and real-world datasets, we demonstrate that our approach effectively uncovers temporal patterns and improves the interpretability of the visualizations. Furthermore, the method improves temporal coherence while preserving the fidelity of the embeddings, providing a robust tool for dynamic data analysis.

A novel ddPCR™ assay for eDNA detection and quantification of Greater Amberjack Seriola dumerilli and three congeners in US waters: challenges and application to fisheries independent surveys

Background Four Seriola species support recreational and commercial fisheries along the U.S. Atlantic Ocean and the Gulf of Mexico, with the S. dumerili Gulf of Mexico stock being overfished for over three decades. The study presented here is part of a fisheries-independent project initiated to determine an absolute abundance of S. dumerili, to expand biological knowledge of the species and to develop novel tools for fisheries management. Environmental DNA (eDNA) tools aimed at the detection and quantification of target species are starting to emerge in support of marine fisheries surveys. Key to progressing the field is Droplet Digital™ PCR (ddPCR™), a highly sensitive technique with advanced multiplexing and direct quantification capabilities that can provide fisheries scientists with improved interpretation of eDNA data. Methods We developed and validated a novel tetraplex ddPCR™ assay able to detect and distinguish between S. dumerili, S. fasciata, S. rivoliana, and S. zonata from seawater eDNA samples. In order to groundtruth ddPCR™ data, and explore its capacity to provide abundance estimates, we compared ddPCR™ detections and quantifications to abundance data inferred from multiple camera (ROV, S-BRUV, chevron trap) and acoustic (VPS array) gears deployed during a fisheries research gear-calibration cruise. Results We demonstrated that with eDNA contamination controls and best practice protocols, it is viable to conduct eDNA research as part of a fisheries survey cruise. eDNA sampling was completed in less time than camera gears (15 min vs 2 h). Both eDNA and camera gears detected the presence of S. dumerili and S. rivoliana at both sites and all sampling days, but not S. fasciata and S. zonata. eDNA concentration data was higher for S. dumerili than S. rivoliana at both sites for all sampling days, in line with abundance patterns obtained from camera gears. The highest correlation (r = 0.97) was obtained between the measures of eDNA between gear deployments and ROV. Discussion Incorporating eDNA in fisheries surveys would not require additional days at sea and could improve precision in fish detection and abundance. eDNA can be a valuable complement to camera gears deployed in geographic areas or seasons with poor visibility conditions, where fish may be present but cannot be confidently identified to the species level. The high correlation obtained between ROV and eDNA data collected between gear deployments adds to a growing number of studies demonstrating the potential of eDNA as an indicator of abundance for fisheries stock assessments. Time-series data from a carefully designed eDNA survey, that estimates relative abundance, could be used as an index of relative abundance for the S. dumerili stock assessment. To achieve this, investment into follow-up studies with increased sample sizes and spatial and temporal replication would be necessary to allow for year-to-year comparisons and validate the robustness of the correlation observed.

CLTNet: A Hybrid Deep Learning Model for Motor Imagery Classification

Background: Brain–computer interface (BCI) technology opens up new avenues for human–machine interaction and rehabilitation by connecting the brain to machines. Electroencephalography (EEG)-based motor imagery (MI) classification is a key component of BCI technology, which is capable of translating neural activity in the brain into commands for controlling external devices. Despite the great potential of BCI technology, the challenges of extracting and decoding brain signals limit its wide application. Methods: To address this challenge, this study proposes a novel hybrid deep learning model, CLTNet, which focuses on solving the feature extraction problem to improve the classification of MI-EEG signals. In the preliminary feature extraction stage, CLTNet uses a convolutional neural network (CNN) to extract time series, channel, and spatial features of EEG signals to obtain important local information. In the deep feature extraction stage, the model combines the long short-term memory (LSTM) network and the Transformer module to capture time-series data and global dependencies in the EEG. The LSTM explains the dynamics of the brain activity, while the Transformer’s self-attention mechanism reveals the global features of the time series. Ultimately, the CLTNet model classifies motor imagery EEG signals through a fully connected layer. Results: The model achieved an average accuracy of 83.02% and a Kappa value of 0.77 on the BCI IV 2a dataset, and 87.11% and a Kappa value of 0.74 on the BCI IV 2b dataset, both of which outperformed the traditional methods. Conclusions: The innovation of the CLTNet model is that it integrates multiple network architectures, which offers a more comprehensive understanding of the characteristics of the EEG signals during motor imagery, providing a more comprehensive perspective and establishing a new benchmark for future research in this area.

Short-Medium-Term Solar Irradiance Forecasting with a CEEMDAN-CNN-ATT-LSTM Hybrid Model Using Meteorological Data

In recent years, the adverse effects of climate change have increased rapidly worldwide, driving countries to transition to clean energy sources such as solar and wind. However, these energies face challenges such as cloud cover, precipitation, wind speed, and temperature, which introduce variability and intermittency in power generation, making integration into the interconnected grid difficult. To achieve this, we present a novel hybrid deep learning model, CEEMDAN-CNN-ATT-LSTM, for short- and medium-term solar irradiance prediction. The model utilizes complete empirical ensemble modal decomposition with adaptive noise (CEEMDAN) to extract intrinsic seasonal patterns in solar irradiance. In addition, it employs a hybrid encoder-decoder framework that combines convolutional neural networks (CNN) to capture spatial relationships between variables, an attention mechanism (ATT) to identify long-term patterns, and a long short-term memory (LSTM) network to capture short-term dependencies in time series data. This model has been validated using meteorological data in a more than 2400 masl region characterized by complex climatic conditions south of Ecuador. It was able to predict irradiance at 1, 6, and 12 h horizons, with a mean absolute error (MAE) of 99.89 W/m2 in winter and 110.13 W/m2 in summer, outperforming the reference methods of this study. These results demonstrate that our model represents progress in contributing to the scientific community in the field of solar energy in environments with high climatic variability and its applicability in real scenarios.

Generative Adversarial Network for Synthesizing Multivariate Time-Series Data in Electric Vehicle Driving Scenarios

Generative Adversarial Network for Synthesizing Multivariate Time-Series Data in Electric Vehicle Driving Scenarios

Detecting and Explaining Long-Term Trends in a Weed Community in a Biennial Cereal–Legume Rotation

The technique of Dynamic Factor Analysis (DFA), which aims to reduce the dimensionality of time-series data, was utilized in order to model the changes over time in eight different long-time-series weeds (26 years) growing in a biennial cereal–legume rotation. The aim of the present study was to determine the existence of long-term trends in a weed community and to identify the factors that determine them. A common trend was extracted that captured the main signal of abundance over time, indicating latent influences affecting all species. Canonical correlation analysis showed strong associations between the common trend and specific weed species, suggesting differential responses to this latent factor. Local (temperature and precipitation) and global weather factors (North Atlantic Oscillation (NAO)) were considered as explanatory variables to explain the common trend. The local weather variables considered did not play a significant role in explaining the commonly observed trend. Conversely, NAO showed a significant relationship with the weed community, indicating its potential role in shaping long-term weed dynamics. DFA was found to be useful for studying the variability in multivariate weed time-series without the need for detailed a priori information on the underlying mechanisms governing weed population dynamics. Overall, this study provided valuable insights into the long-term drivers of weed dynamics and set the stage for future research in this area.

Research on Automatic Identification of Power Big Data Anomalies Based on Improved Decision Tree

With the advancement of smart grid technology, the issue of power system network security has become increasingly critical. To fully utilize the power grid’s vast data resources and enhance the efficiency of anomaly detection, this paper proposes an improved decision tree (DT)-based automatic identification approach for anomalies in electric power big data. The method employs six-dimensional features extracted from the dimensions of volatility, trend, and variability to characterize the time series of power data. These features are integrated into a hybrid DT-SVM-LSTM framework, combining the strengths of DTs, support vector machines, and long short-term memory networks. Experimental results demonstrate that the proposed method achieves an accuracy of 96.8%, a precision of 95.3%, a recall of 94.8%, and an F1-score of 95.0%, outperforming several state-of-the-art methods cited in the literature. Moreover, the approach exhibits strong robustness to noise, maintaining high detection accuracy even under low signal-to-noise ratio conditions. These findings highlight the effectiveness of the method in efficiently detecting anomalies and addressing noise interference.

Long-Term Forecasting of Solar Irradiation in Riyadh, Saudi Arabia, Using Machine Learning Techniques

Forecasting of time series data presents some challenges because the data’s nature is complex and therefore difficult to accurately forecast. This study presents the design and development of a novel forecasting system that integrates efficient data processing techniques with advanced machine learning algorithms to improve time series forecasting across the sustainability domain. Specifically, this study focuses on solar irradiation forecasting in Riyadh, Saudi Arabia. Efficient and accurate forecasts of solar irradiation are important for optimizing power production and its smooth integration into the utility grid. This advancement supports Saudi Arabia in Vision 2030, which aims to generate and utilize renewable energy sources to drive sustainable development. Therefore, the proposed forecasting system has been developed to the parameters characteristic of the Riyadh region environment, including high solar intensity, dust storms, and unpredictable weather conditions. After the cleaning and filtering process, the filtered dataset was pre-processed using the standardization method. Then, the Discrete Wavelet Transform (DWT) technique has been applied to extract the features of the pre-processed data. Next, the extracted features of the solar dataset have been split into three subsets: train, test, and forecast. Finally, two different machine learning techniques have been utilized for the forecasting process: Support Vector Machine (SVM) and Gaussian Process (GP) techniques. The proposed forecasting system has been evaluated across different time horizons: one-day, five-day, ten-day, and fifteen-day ahead. Comprehensive evaluation metrics were calculated including accuracy, stability, and generalizability measures. The study outcomes present the proposed forecasting system which provides a more robust and adaptable solution for time-series long-term forecasting and complex patterns of solar irradiation in Riyadh, Saudi Arabia.

MM-iTransformer: A Multimodal Approach to Economic Time Series Forecasting with Textual Data

This paper introduces a novel multimodal framework for economic time series forecasting, integrating textual information with historical price data to enhance predictive accuracy. The proposed method employs a multi-head attention mechanism to dynamically align textual embeddings with temporal price data, capturing previously unrecognized cross-modal dependencies and enhancing the model’s ability to interpret event-driven market dynamics. This enables the framework to model complex market behaviors in a unified and effective manner. Experimental results across multiple financial datasets, including the foreign exchange (Forex) and Gold-price datasets, demonstrate that incorporating textual information significantly enhances forecasting accuracy. Compared to models relying solely on historical price data, the proposed framework achieves a substantial reduction in Mean Squared Error (MSE) loss, with improvements of up to 26.79%. This highlights the effectiveness of leveraging textual inputs alongside structured time series data in capturing complex market dynamics and improving predictive performance.