Search

Multi-channel Electroencephalogram (EEG) based emotion recognition is focused on several analysis of frequency bands of the acquired signals. In this paper, spectral properties appeared on five EEG bands (δ , θ , α , β , γ ) and gated transformer network (GTN) based emotion recognition using EEG signal are proposed. Spectral energies and differential entropies of 62-channel signals are converted to 3D (sequence-channel-trial) form to feed the GTN. The GTN with enhanced gated two tower based transformer architecture is fed by 3D sequences extracted from SEED and SEED-IV emotional datasets. 15 participants' states in session 1–3 are evaluated using the proposed GTN based sequence classification, and the results are repeated by 3 × shuffling. Totally, 135 times training and testing are performed on each dataset, and the results are presented. The proposed GTN model achieves mean accuracy rates of 98.82% on the SEED dataset and 96.77% on the SEED-IV dataset for three and four emotional state recognition tasks, respectively. The proposed emotion recognition model can be employed as a promising approach for EEG emotion recognition. [ABSTRACT FROM AUTHOR]

In this paper, we propose a correlation‐based test for the evaluation of two competing forecasts. Under the null hypothesis of equal correlations with the target variable, we derive the asymptotic distribution of our test using the Delta method. This null hypothesis is not necessarily equivalent to the null of equal Mean Squared Prediction Errors (MSPE). Specifically, it might be the case that the forecast displaying the lowest MSPE also exhibits the lowest correlation with the target variable: this is known as "The MSPE paradox." In this sense, our approach should be seen as complementary to traditional tests of equality in MSPE. Monte Carlo simulations indicate that our test has good size and power. Finally, we illustrate the use of our test in an empirical exercise in which we compare two different inflation forecasts for a sample of OECD economies. We find more rejections of the null of equal correlations than rejections of the null of equality in MSPE. [ABSTRACT FROM AUTHOR]

Monitoring active volcanoes is necessary to analyze their current status to pose a mitigation hazard. Mount Melbourne is an active volcano that has erupted in the past, and future eruptions are possible. This condition could threaten future eruptions, particularly near scientific bases. Jang Bogo, a South Korean research station, is located only 30 km from the summit and could be affected by significant ash fallout in case of an explosive eruption. This condition leads to the necessity of observing Mount Melbourne's activity frequently. This study used Sentinel-1 SAR data acquired from 2017 to 2024 to monitor the volcanic activity of Mount Melbourne by utilizing InSAR multitemporal time-series analysis implementing the improved combined scatterers interferometry with optimized point scatterers (ICOPS) method. The ICOPS method combined persistent scatterer (PS) and distributed scatterer (DS) with measurement point (MP) optimization based on convolutional neural network (CNN) and optimized hot spot analysis (OHSA). The ICOPS measurement results maintain reliable MP along the Mount Melbourne summit and around Jang Bogo station. The absence of GPS stations around these two areas makes it difficult to validate the result with the ground truth measurement, so the comparison with another method, small baseline (SBAS) measurement, is made to evaluate the reliability of the ICOPS measurement points. The comparison between the MP from ICOPS and the SBAS methods shows a good correlation with R2 of about 0.8134 in the Melbourne area and 0.8678 in the Jang Bogo area. The selected time-series plot around the summit of Mount Melbourne and the Jang Bogo area shows a stable trend of surface deformation. Thus, a total accumulated deformation of around 0.82 cm and an average deformation of about 0.10 cm/year was found around Mount Melbourne. Meanwhile, the Jang Bogo area exhibits a total deformation of about 0.15 cm with an average deformation of about 0.02. Overall, this research is a preliminary study of the ability of the ICOPS algorithm to monitor volcanic activity in snow-covered areas. [ABSTRACT FROM AUTHOR]

Clustering algorithms such as k-means and agglomerative hierarchical clustering (HCA) may provide a unique opportunity to analyze time-series kinematic data. Here we present an approach for determining number of clusters and which clustering algorithm to use on time-series lumbar and pelvis kinematic data. Cluster evaluation measures such as silhouette coefficient, elbow method, Dunn Index, and gap statistic were used to evaluate the quality of decision making. The result show that multiple clustering evaluation methods should be used to determine the ideal number of clusters and algorithm suitable for clustering time-series data for each dataset being analyzed. [ABSTRACT FROM AUTHOR]

The Tigris and Euphrates Basin (TEB) encompasses a wide area, and due to its different geographical and political conditions, each environmental factor in different conditions has different effects on the process of surface water changes. Accordingly, in this research, we aim to evaluate the trend of surface water changes in this basin in 2001-2021 by using the time series of 16 different parameters and the products available on the Google Earth Engine (GEE) platform. Based on the findings, the general trend of surface water changes is increasing, and the water area has reached 8605.9 km2 in 2001 to 10021.8 km2 in 2021. Nevertheless, the spatial-temporal changes of water have been different because the extent of lakes and wetlands in the southern areas of the basin has decreased drastically. On the contrary, it has increased upstream of the basin due to the expansion of various dams and channels. In addition, our findings indicated a high correlation between climatic variables, especially evapotranspiration, and temperature, with temporal changes in water in the region. Thus, the impact of global climate changes on the hydrology and environment of the basin highlights the importance and high sensitivity of the major lakes in the region, such as Razazeh, Tharthar, Hamrin, and Habbaniyah, to climate changes. The present research results may be used to assess surface water in other regions and provide valuable information for the planning and management of global surface water resources. [ABSTRACT FROM AUTHOR]

In this work, we explored the feasibility of using a transformer-based time-series forecasting architecture, known as the Temporal Fusion Transformer (TFT), for anomaly detection using threaded track data from the MITRE Corporation's Transportation Data Platform (TDP) and digital flight data. The TFT architecture has the flexibility to include both time-varying multivariate data and categorical data from multimodal data sources and conduct single-output or multi-output predictions. For anomaly detection, rather than training a TFT model to predict the outcomes of specific aviation safety events, we train a TFT model to learn nominal behavior. Any significant deviation of the TFT model's future horizon forecast for the output flight parameters of interest from the observed time-series data is considered an anomaly when conducting evaluations. For proof-of-concept demonstrations, we used an unstable approach (UA) as the anomaly event. This type of anomaly detection approach with nominal behavior learning can be used to develop flight analytics to identify emerging safety hazards in historical flight data and has the potential to be used as an on-board early warning system to assist pilots during flight. [ABSTRACT FROM AUTHOR]

We propose a novel methodology for forecasting spatio-temporal data using supervised semi-nonnegative matrix factorization (SSNMF) with frequency regularization. Matrix factorization is employed to decompose spatio-temporal data into spatial and temporal components. To improve clarity in the temporal patterns, we introduce a nonnegativity constraint on the time domain along with regularization in the frequency domain. Specifically, regularization in the frequency domain involves selecting features in the frequency space, making an interpretation in the frequency domain more convenient. We propose two methods in the frequency domain: soft and hard regularizations, and provide convergence guarantees to first-order stationary points of the corresponding constrained optimization problem. While our primary motivation stems from geophysical data analysis based on GRACE (Gravity Recovery and Climate Experiment) data, our methodology has the potential for wider application. Consequently, when applying our methodology to GRACE data, we find that the results with the proposed methodology are comparable to previous research in the field of geophysical sciences but offer clearer interpretability. [ABSTRACT FROM AUTHOR]

As engine monitoring data has become more complex with an increasing number of sensors, fault prediction based on artificial intelligence (AI) has emerged. Existing fault prediction models using AI significantly improve the accuracy of predictions by effectively handling such complex data, but at the same time, the problem arises that the AI-based models cannot explain the rationale of their predictions to users. To address this issue, we propose a time-series explanatory fault prediction framework to provide an explainability even when using AI-based fault prediction models. It consists of a data feature reduction process, a fault prediction model training process using long short-term memory, and an interpretation process of the fault prediction model via an explainable AI method. In particular, the proposed framework can explain a fault prediction based on time-series data. Therefore, it indicates which part of the data was significant for the fault prediction not only in terms of sensor type but also in terms of time. Through extensive experiments, we evaluate the proposed framework using various fault data by comparing the prediction performance of fault prediction and by assessing how well the main pre-symptoms of the fault are extracted when predicting a fault. [ABSTRACT FROM AUTHOR]

Understanding patterns of species diversity is crucial for ecological research and conservation, and this understanding may be improved by studying patterns in the two components of species diversity, species richness and evenness of abundance of species. Variation in species richness and evenness has previously been linked to variation in total abundance of communities as well as productivity gradients. Exploring both components of species diversity is essential because these components could be unrelated or driven by different mechanisms. The aim of this study was to investigate the relationship between species richness and evenness in European bird communities along an extensive latitudinal gradient. We examined their relationships with latitude and Net Primary Productivity, which determines energy and matter availability for heterotrophs, as well as their responses to territory densities (i.e. the number of territories per area) and community biomass (i.e. the bird biomass per area). We applied a multivariate Poisson log‐normal distribution to unique long‐term, high‐quality time‐series data, allowing us to estimate species richness of the community as well as the variance of this distribution, which acts as an inverse measure of evenness. Evenness in the distribution of abundance of species in the community was independent of species richness. Species richness increased with increasing community biomass, as well as with increasing density. Since both measures of abundance were explained by NPP, species richness was partially explained by energy‐diversity theory (i.e. the more energy, the more species sustained by the ecosystem). However, species richness did not increase linearly with NPP but rather showed a unimodal relationship. Evenness was not explained either by productivity nor by any of the aspects of community abundance. This study highlights the importance of considering both richness and evenness to gain a better understanding of variation in species diversity. We encourage the study of both components of species diversity in future studies, as well as use of simulation studies to verify observed patterns between richness and evenness. [ABSTRACT FROM AUTHOR]

Accurate resource utilization estimation is crucial for efficient resource allocation, capacity planning, and cost optimization in cloud systems. In the past, several artificial intelligence, machine learning, and deep learning-based techniques have been evolved to forecast cloud cluster workload. Despite the abundance of available techniques, most existing techniques fail to achieve the desired prediction efficiency and generalization capability. They are computationally inefficient in accurately determining resource utilization at a machine-level granularity. The current research proposes a computationally less-expensive hybrid approach combining cluster analysis and deep neural learning with transfer learning to estimate the machine-level workload. The method implements clustering to identify the similarity patterns among the non-linear usage profiles of machines present in the input dataset. Subsequently, the generalized deep neural learning models are developed considering only a sample dataset belonging to each identified cluster. Lastly, the concept of transfer learning is deployed using pre-trained generalized models to estimate the workload for all remaining machines relating to the clusters. The performance validation of the proposed approach is carried out on the real-world traces dataset of the google cluster. The comparative evaluation of the proposed approach with benchmark approaches verifies the achieved performance benefits and accuracy. [ABSTRACT FROM AUTHOR]

Accurate cloud and shadow detection is a crucial prerequisite for optical remote sensing image analysis and application. Multi-temporal-based cloud and shadow detection methods are a preferable choice to detect clouds in complex scenes (e.g., thin clouds, broken clouds and clouds with interference from artificial surfaces with high reflectivity). However, such methods commonly require cloud-free reference images, and this may be difficult to achieve in time-series data since clouds are often prevalent and of varying spatial distribution in optical remote sensing images. Furthermore, current multi-temporal-based methods have limited feature extraction capability and rely heavily on prior assumptions. To address these issues, this paper proposes a Siamese network (Siamnet) for cloud and shadow detection based on Time-Series cloudy Images, namely TSI-Siamnet, which consists of two steps: 1) low-rank and sparse component decomposition of time-series cloudy images is conducted to construct a composite reference image to cope with dynamic changes in the cloud distribution in time-series images; 2) an extended Siamnet with optimal difference calculation module (DM) and multi-scale difference features fusion module (MDFM) is constructed to extract reliable disparity features and alleviate semantic information feature dilution during the decoder part. TSI-Siamnet was tested extensively on seven land cover types in the well-known Landsat 8 Biome dataset. Compared to six state-of-the-art methods (including four deep learning-based methods and two classical non-deep learning-based methods), TSI-Siamnet produced the best performance with an overall accuracy of 95.05% and MIoU of 84.37%. In three more challenging experiments, TSI-Siamnet showed enhanced detection of thin and broken clouds and greater anti-interference to highly reflective surfaces. TSI-Siamnet provides a novel strategy to explore comprehensively the valid information in time-series cloudy images and integrate the extracted spectral-spatial–temporal features for reliable cloud and shadow detection. [ABSTRACT FROM AUTHOR]

The hadopelagic environment remains highly understudied due to the inherent difficulties in sampling at these depths. The use of sediment environmental DNA (eDNA) can overcome some of these restrictions as settled and preserved DNA represent an archive of the biological communities. We use sediment eDNA to assess changes in the community within one of the world's most productive open‐ocean ecosystems: the Atacama Trench. The ecosystems around the Atacama Trench have been intensively fished and are affected by climate oscillations, but the understanding of potential impacts on the marine community is limited. We sampled five sites using sediment cores at water depths from 2400 to ~8000 m. The chronologies of the sedimentary record were determined using 210Pbex. Environmental DNA was extracted from core slices and metabarcoding was used to identify the eukaryote community using two separate primer pairs for different sections of the 18S rRNA gene (V9 and V7) effectively targeting pelagic taxa. The reconstructed communities were similar among markers and mainly composed of chordates and members of the Chromista kingdom. Alpha diversity was estimated for all sites in intervals of 15 years (from 1842 to 2018), showing a severe drop in biodiversity from 1970 to 1985 that aligns with one of the strongest known El Niño events and extensive fishing efforts during the time. We find a direct impact of sea surface temperature on the community composition over time. Fish and cnidarian read abundance was examined separately to determine whether fishing had a direct impact, but no direct relation was found. These results demonstrate that sediment eDNA can be a valuable emerging tool providing insight in historical perspectives on ecosystem developments. This study constitutes an important step toward an improved understanding of the importance of environmental and anthropogenic drivers in affecting open and deep ocean communities. [ABSTRACT FROM AUTHOR]

This study aimed to improve the predictive accuracy of the Braden assessment for pressure injury risk in skilled nursing facilities (SNFs) by incorporating real‐world data and training a survival model. A comprehensive analysis of 126 384 SNF stays and 62 253 in‐house pressure injuries was conducted using a large calibrated wound database. This study employed a time‐varying Cox Proportional Hazards model, focusing on variations in Braden scores, demographic data and the history of pressure injuries. Feature selection was executed through a forward‐backward process to identify significant predictive factors. The study found that sensory and moisture Braden subscores were minimally contributive and were consequently discarded. The most significant predictors of increased pressure injury risk were identified as a recent (within 21 days) decrease in Braden score, low subscores in nutrition, friction and activity, and a history of pressure injuries. The model demonstrated a 10.4% increase in predictive accuracy compared with traditional Braden scores, indicating a significant improvement. The study suggests that disaggregating Braden scores and incorporating detailed wound histories and demographic data can substantially enhance the accuracy of pressure injury risk assessments in SNFs. This approach aligns with the evolving trend towards more personalized and detailed patient care. These findings propose a new direction in pressure injury risk assessment, potentially leading to more effective and individualized care strategies in SNFs. The study highlights the value of large‐scale data in wound care, suggesting its potential to enhance quantitative approaches for pressure injury risk assessment and supporting more accurate, data‐driven clinical decision‐making. [ABSTRACT FROM AUTHOR]

Abstract Background The relationship between human gut microbiota and high-altitude hypoxia acclimatization remains highly controversial. This stems primarily from uncertainties regarding both the potential temporal changes in the microbiota under such conditions and the existence of any dominant or core bacteria that may assist in host acclimatization. Results To address these issues, and to control for variables commonly present in previous studies which significantly impact the results obtained, namely genetic background, ethnicity, lifestyle, and diet, we conducted a 108-day longitudinal study on the same cohort comprising 45 healthy Han adults who traveled from lowland Chongqing, 243 masl, to high-altitude plateau Lhasa, Xizang, 3658 masl, and back. Using shotgun metagenomic profiling, we study temporal changes in gut microbiota composition at different timepoints. The results show a significant reduction in the species and functional diversity of the gut microbiota, along with a marked increase in functional redundancy. These changes are primarily driven by the overgrowth of Blautia A, a genus that is also abundant in six independent Han cohorts with long-term duration in lower hypoxia environment in Shigatse, Xizang, at 4700 masl. Further animal experiments indicate that Blautia A-fed mice exhibit enhanced intestinal health and a better acclimatization phenotype to sustained hypoxic stress. Conclusions Our study underscores the importance of Blautia A species in the gut microbiota’s rapid response to high-altitude hypoxia and its potential role in maintaining intestinal health and aiding host adaptation to extreme environments, likely via anti-inflammation and intestinal barrier protection.

In this article, we address the potential effects of construction activities around a permanent Global Navigation Satellite System (GNSS) station. The study is based on a consistent time series of GNSS observations collected over a span of almost 13 years and comprises an analysis in both observation and coordinate domains. The analysis in the observation domain revealed a significant increase of the code multipath for the C1C signal of GPS and GLONASS. In terms of standard deviations the values increased by 37 and 31%, respectively. No significant increase was observed for other signals. Interestingly, the analysis showed that the Galileo E5 AltBOC signal outperformed the other signals in distinguishing direct from indirect signals. In the coordinate domain, the relative GNSS solution results demonstrated an increase in root mean square error for the eastern component during the same period. Noteworthy, the observed increase in both domains coincided with the period of the highest level of construction activity around the GNSS station. These findings suggest that the construction activity near the station impacted the quality of GNSS observations and that the Galileo E5 AltBOC signal was better at compensating for local changes than other signals.

Accurate, consistent, and high-resolution Land Use & Cover (LUC) information is fundamental for effectively monitoring landscape dynamics and better apprehending drivers, pressures, state, and impacts on land systems. Nevertheless, the availability of such national products with high thematic accuracy is still limited and consequently researchers and policymakers are constrained to work with data that do not necessarily reflect on-the-ground realities impending to correctly capture details of landscape features as well as limiting the identification and quantification of drivers and rate of change. Hereafter, we took advantage of the Switzerland’s official LUC statistical sampling survey and dense time-series of Sentinel-2 data, combining them with Machine and Deep Learning methods to produce an accurate and high spatial resolution land cover map over the Lake Geneva region. Findings suggest that time-first approach is a valuable alternative to space-first approaches, accounting for the intra-annual variability of classes, hence improving classification performances. Results demonstrate that Deep Learning methods outperform more traditional Machine Learning ones such as Random Forest, providing more accurate predictions with lower uncertainty. The produced land cover map has a high accuracy, an improved spatial resolution, while at the same time preserving the statistical significance (i.e. class proportion) of the official national dataset. This work paves the way towards the objective to produce a yearly high resolution land cover map of Switzerland and potentially implement a continuous land change monitoring capability. However further work is required to properly address challenges such as the need for increased temporal resolution for LUC information or the quality of training samples.

Abstract Background Extracellular vesicles (EVs) hold the potential for elucidating the pathogenesis of amyotrophic lateral sclerosis (ALS) and serve as biomarkers. Notably, the comparative and longitudinal alterations in the protein profiles of EVs in serum (sEVs) and cerebrospinal fluid (CSF; cEVs) of sporadic ALS (SALS) patients remain uncharted. Ropinirole hydrochloride (ROPI; dopamine D2 receptor [D2R] agonist), a new anti-ALS drug candidate identified through induced pluripotent stem cell (iPSC)-based drug discovery, has been suggested to inhibit ALS disease progression in the Ropinirole Hydrochloride Remedy for Amyotrophic Lateral Sclerosis (ROPALS) trial, but its mechanism of action is not well understood. Therefore, we tried to reveal longitudinal changes with disease progression and the effects of ROPI on protein profiles of EVs. Methods We collected serum and CSF at fixed intervals from ten controls and from 20 SALS patients participating in the ROPALS trial. Comprehensive proteomic analysis of EVs, extracted from these samples, was conducted using liquid chromatography/mass spectrometer (LC/MS). Furthermore, we generated iPSC-derived astrocytes (iPasts) and performed RNA sequencing on astrocytes with or without ROPI treatment. Results The findings revealed notable disparities yet high congruity in sEVs and cEVs protein profiles concerning disease status, time and ROPI administration. In SALS, both sEVs and cEVs presented elevated levels of inflammation-related proteins but reduced levels associated with unfolded protein response (UPR). These results mirrored the longitudinal changes after disease onset and correlated with the revised ALS Functional Rating Scale (ALSFRS-R) at sampling time, suggesting a link to the onset and progression of SALS. ROPI appeared to counteract these changes, attenuating inflammation-related protein levels and boosting those tied to UPR in SALS, proposing an anti-ALS impact on EV protein profiles. Reverse translational research using iPasts indicated that these changes may partly reflect the DRD2-dependent neuroinflammatory inhibitory effects of ROPI. We have also identified biomarkers that predict diagnosis and disease progression by machine learning-driven biomarker search. Conclusions Despite the limited sample size, this study pioneers in reporting time-series proteomic alterations in serum and CSF EVs from SALS patients, offering comprehensive insights into SALS pathogenesis, ROPI-induced changes, and potential prognostic and diagnostic biomarkers.

Background and aims Design, setting and participants Measurements Findings Conclusions Public health concerns regarding pregnant women’s health after the enactment of the Cannabis Act in Canada (CAC) (a law that allowed non‐medical cannabis use), and the potential impact of the COVID‐19 pandemic, call for a contemporary assessment of these two events. Our study measured associations between the CAC, the COVID‐19 pandemic and the monthly prevalence rates of cannabis‐, all drug‐ and alcohol‐related diagnosed disorders among pregnant women in the province of Quebec.This was a quasi‐experimental design applying an interrupted time‐series methodology in the province of Quebec, Canada. The participants were pregnant women aged 15–49 years, between January 2010 and July 2022.Administrative health data from the Québec Integrated Chronic Disease Surveillance System were used to classify pregnant women according to cannabis‐, all drug (excluding cannabis)‐ and alcohol‐related disorders. The CAC (October 2018) and the COVID‐19 pandemic (April 2020) were evaluated as (1) slope changes and (2) level changes. Cannabis‐, all drug (excluding cannabis)‐ and alcohol‐related disorders were measured by total monthly age‐standardized monthly prevalence rate of each disorder for pregnant women aged 15–49 years.Before the CAC, the prevalence rate of cannabis‐related diagnosed disorders significantly increased each month by 0.5% [95% confidence interval (CI) = 0.3–0.6] in the pregnant population. After the CAC, there were significant increases of 24% (95% CI = 1–53) of cannabis‐related diagnosed disorders. No significant changes were observed for all drug (excluding cannabis)‐ and alcohol‐related diagnosed disorders associated with the CAC. A non‐significant decrease of 20% (95% CI = −38 to 3) was observed during the COVID‐19 pandemic in alcohol‐related disorders.The monthly incidence rates of diagnosed cannabis‐related disorders in pregnant women in Quebec increased significantly following the enactment of the Cannabis Act in Canada. Diagnoses of all drug (excluding cannabis)‐ and alcohol‐related disorders remained relatively stable. [ABSTRACT FROM AUTHOR]

Multiple evidence has supported that air pollution exposure has detrimental effects on the cardiovascular and respiratory systems. However, most investigations focus on the general population, with limited research conducted on medically insured populations. To address this gap, the current research was designed to examine the acute effects of inhalable particulate matter (PM2.5 and PM10), nitrogen dioxide (NO2), ground-level ozone (O3), and sulfur dioxide (SO2) on the incidence of upper respiratory tract infections (URTI), utilizing medical insurance data in Wuhan, China. Data on URTI were collected from the China Medical Insurance Basic Database for Wuhan covering the period from 2014 to 2018, while air pollutant data was gathered from ten national monitoring stations situated in Wuhan city. Statistical analysis was performed using generalized additive models for quasi-Poisson distribution with a log link function. The analysis indicated that except for ozone, higher exposure to four other pollutants (NO2, SO2, PM2.5, and PM10) were significantly linked to an elevated risk of URTI, particularly during the previous 0–3 days and previous 0–4 days. Additionally, NO2 and SO2 were found to be positively linked with laryngitis. Furthermore, the effects of air pollutants on the risk of URTI were more pronounced during cold seasons than hot seasons. Notably, females and the employed population were more susceptible to infection than males and non-employed individuals. Our findings gave solid proof of the link between ambient air pollution exposure and the risk of URTI in medically insured populations. [ABSTRACT FROM AUTHOR]

• Landsat archive presents unique opportunities for multitemporal analysis. • Specific month selection in time series effects classification accuracy. • Benefit of increased month observations varied among climatic regions and classes. • Grass/shrub and cultivated classes saw highest improvements. • Implementation scenario showed F1 improvements as high as 10% The Landsat archive having consistent revisit times, near global extent and extensive multi-decadal temporal coverage offers a unique opportunity for land cover land use product generation. Along with this vast volume of freely available data, new classification methods based on deep learning have improved modeling capabilities. This manuscript investigates the effect of intra-annual Landsat scene availability in the accuracy of land cover land use classification in the conterminous United States. More specifically, we seek to quantify the effect of: i) increased monthly scene availability, and ii) specific months that may result in higher classification accuracy across different classes. Identifying specific months with comparable classification accuracy to the entire time series could offer significant computational gains for large-scale mapping. Our experiment incorporated deep learning classifiers and a wide range of reference data across the continental United States. Results were contrasted between five large U.S. climatic regions to further differentiate this intra-annual effect. Our findings indicate that the total number of months can have a highly variable effect in the classification accuracy ranging from minor (a few percentage points in terms of class F1 accuracy) to extremely beneficial (approaching 50% F1 improvement moving from four to twelve month observations). The benefit of increased month observations varied among climatic regions and classes: when all climate regions were combined, the grass/shrub and cultivated classes improved their F1 accuracy up to 30%, while the water class saw the least improvement of about 5%, partially due to its limited room for improvement. The effect of specific month combinations was also examined, where the total number of months was kept constant and the included months varied. The difference between the best month combination and the median combination value was estimated to be as high as about 30% for the four monthly observations scenario and the grass/shrub class. Further validation of the month selection importance comes from an example implementation scenario where F1 improvements can be as high as 10%. Our work demonstrated that month selection may offer such benefits that in some classes and climatic regions this time selection optimization is an inevitable choice due to large accuracy improvements. Also, the potential data reduction with targeted month selection would be particularly appealing to large-scale classification tasks. Due to the large extent of the climatic regions further studies are needed to quantify a more localized effect along with explanation of potential drivers. [ABSTRACT FROM AUTHOR]

The importance of forecasting crop yields in agriculture cannot be overstated. The effects of yield forecasting are observed in all the aspects of the supply chain from staffing to supplier demand, food waste, and other business decisions. However, the process is often inaccurate and far from perfect. This paper explores the potential of using expert forecasts to enhance the crop yield predictions of our global-to-local XGBoost machine learning system. Additionally, it investigates the ERA5 climate model's viability as an alternative data source for crop yield forecasting in the absence of on-farm weather data. We find that, by combining both the expert's pre-season forecasts and the ERA5 climate model with the machine learning model, we can—in most cases—obtain better forecasts that outperform the growers' pre-season forecasts and the machine learning-only models. Our expert-informed model attains yield forecasts for 4 weeks ahead with an average RMSE of 0.0855 across all the plots and an RMSE of 0.0872 with the ERA5 climate data included. [ABSTRACT FROM AUTHOR]

As big data technologies for IoT services develop, cross-service distributed learning techniques of multivariate deep learning models on IoT time-series data collected from various sources are becoming important. Vertical federated deep learning (VFDL) is used for cross-service distributed learning for multivariate IoT time-series deep learning models. Existing VFDL methods with reasonable performance require a large communication amount. On the other hand, existing communication-efficient VFDL methods have relatively low performance. We propose TT-VFDL-SIM, which can achieve improved performance over centralized training or existing VFDL methods in a communication-efficient manner. TT-VFDL-SIM derives partial tasks from the target task and applies transfer learning to them. In our task-driven transfer approach for the design of TT-VFDL-SIM, the SIM Partial Training mechanism contributes to performance improvement by introducing similar feature spaces in various ways. TT-VFDL-SIM was more communication-efficient than existing VFDL methods and achieved an average of 0.00153 improved MSE and 7.98% improved accuracy than centralized training or existing VFDL methods. [ABSTRACT FROM AUTHOR]

In this study, employing the IEEE Xplore database as the data source, articles on different topics (keywords) and their usage data generated from January 2011 to December 2020 were collected and analyzed. The study examined the temporal relationships between these usage data and publication counts at the topic level via Granger causality analysis. The study found that almost 80% of the topics exhibit significant usage-publication interactions from a time-series perspective, with varying time lag lengths depending on the direction of the Granger causality results. Topics that present bidirectional Granger causality show longer time lag lengths than those exhibiting unidirectional causality. Additionally, the study found that the direction of the unidirectional Granger causality was influenced by the significance of a topic. Topics with a greater preference for article usage as the Granger cause of publication counts were deemed more important. The findings' reliability was confirmed by varying the maximum lag period. This study provides strong support for using usage data to identify hot topics of research. [ABSTRACT FROM AUTHOR]

Global food security is being threatened by the reduction of high‐quality cropland, extreme weather events, and the uncertainty of food supply chains. The globalization of agricultural trade has elevated the diversification of non‐grain production (NGP) on cultivated land to a prominent strategy for poverty alleviation in numerous developing nations. Its rapid expansion has engendered a multitude of deleterious consequences on both food security and ecological stability. NGP in China is becoming very common in the process of rapid urbanization, threatening national food security. To better understand the causal mechanisms and enable governments to balance food security and rural development, it is crucial to have a clear understanding of the spatiotemporal dynamics of NGP using remote sensing. Yet knowledge gaps remain concerning how to use remote sensing to track human‐dominated or ‐induced long‐term cultivated land changes. Our study proposed a method for detecting the spatiotemporal evolution of NGP based on Landsat time‐series data under the Google Earth Engine platform. This approach was proposed by (1) obtaining the union of cultivated lands from multiple landcover products to minimize the cultivated land omission, (2) constructing multi‐index dynamic trend rules for 3 representative types of NGP and obtaining results at the pixel level, while adopting the continuous change detection and classification algorithm to Landsat time series (1986–2022) to determine when the most recent change occurred, (3) minimizing the noise by object‐oriented land use–land cover classification and mode filter approaches, and (4) mapping the spatiotemporal distribution of NGP. The proposed methodology was tested in Jiashan, located in Zhejiang Province (eastern China), where NGP is widespread. We achieved a high overall accuracy of 95.67% for NGP type detection and an overall accuracy of 85.26% for change detection of time. The results indicated a continued increasing pattern of NGP in Jiashan from 1986 to 2022, with the cumulative percentage of NGP increasing from 0.02% to 20.69%. This study highlights the utilization of time‐series data to document essential NGP information for evaluating food security in China and the method is well‐suited for large‐scale mapping due to its automatic manner. [ABSTRACT FROM AUTHOR]

COVID-19 is a deadly infection that causes severe acute respiratory syndrome. Although not particularly spreading rapidly as before due to the introduction of vaccines and other measures, its effect still portends grave danger to human lives in Nigeria and other countries. This study aimed to model and forecast Nigeria's COVID-19 (new) trend of confirmed cases, discharged (recovery) cases and deaths and also to examine the pattern of the infection and survival rate in the face of vaccine introduction. The Box-Jenkins methodology was employed in this study to model and forecast COVID-19 confirmed cases, discharged cases and deaths. The data used for this study was secondary data of weekly confirmed cases, recoveries (discharged) and deaths extracted from the weekly publication of the Nigeria Centre for Disease Control (NCDC). The mean survival rate of COVID-19 was found to be 0.7765 and the three series were found to be stationary after differencing. Also, from an array of candidate models obtained through Autocorrelation Function (ACF) and partial autocorrelation function (PACF) plots, the best-fitted models selected based on minimum Akaike Information Criterion (AIC) and Schwarz Information Criterion (SIC) were found to be ARIMA (4,1,0), (3,1,0) and (7,3,1) for newly confirmed cases, discharged cases and death, respectively. This implied that these models were adequate for forecasting future rates of infection, recovery and death as further diagnostic tests showed that the ARIMA models were the perfect fit for the three cases (since P>0.05). Finally, a 29-week out of sample forecast showed a steep downward trend in the three cases and in particular a drastic decline to zero of future COVID-19 deaths. Based on these results, it was recommended that existing vaccination strategies should be expanded to achieve near-zero new COVID-19 cases and deaths. [ABSTRACT FROM AUTHOR]

Solar photovoltaic (PV) energy, with its clean, local, and renewable features, is an effective complement to traditional energy sources today. However, the photovoltaic power system is highly weather-dependent and therefore has unstable and intermittent characteristics. Despite the negative impact of these features on solar sources, the increase in worldwide installed PV capacity has made solar energy prediction an important research topic. This study compares three encoder-decoder (ED) networks for day-ahead solar PV energy prediction: Long Short-Term Memory ED (LSTM-ED), Convolutional LSTM ED (Conv-LSTM-ED), and Convolutional Neural Network and LSTM ED (CNN-LSTM-ED). The models are tested using 1741-day-long datasets from 26 PV panels in Istanbul, Turkey, considering both power and energy output of the panels and meteorological features. The results show that the Conv-LSTM-ED with 50 iterations is the most successful model, achieving an average prediction score of up to 0.88 over R-square (R2). Evaluation of the iteration counts' effect reveals that the Conv-LSTM-ED with 50 iterations also yields the lowest Root Mean Squared Error (RMSE) and Mean Absolute Error (MAE) values, confirming its success. In addition, the fitness and effectiveness of the models are evaluated, with the Conv-LSTM-ED achieving the lowest Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) values for each iteration. The findings of this work can help researchers build the best data-driven methods for forecasting PV solar energy based on PV features and meteorological features. [ABSTRACT FROM AUTHOR]

Dengue, a disease with multifactorial determinants, is linked to population susceptibility to circulating viruses and the extent of vector infestation. This study aimed to analyze the temporal trends of dengue cases and deaths in Belo Horizonte, Minas Gerais, Brazil, from 2007 to 2020. Data from the Notifiable Diseases Information System (Sinan) were utilized for the investigation. To assess the disease's progression over the study period and predict its future incidence, time series analyses were conducted using a generalized additive model (GAM) and a seasonal autoregressive integrated moving average (SARIMA) model. Over the study period, a total of 463,566 dengue cases and 125 deaths were reported. Notably, there was an increase in severe cases and deaths, marking hyperendemics characterized by simultaneous virus circulation (79.17% in 2016–50% in 2019). The generalized additive model revealed a non-linear pattern with epidemic peaks in 2010, 2013, 2016, and 2019, indicating an explosive pattern of dengue incidence. The SARIMA (3,1,1) (0,0,0)12 model was validated for each year (2015 to 2019). Comparing the actual and predicted numbers of dengue cases, the model demonstrated its effectiveness for predicting cases in the municipality. The rising number of dengue cases emphasizes the importance of vector surveillance and control. Enhanced models and predictions by local health services will aid in anticipating necessary control measures to combat future epidemics. [ABSTRACT FROM AUTHOR]

Railway infrastructure maintenance faces a challenge posed by the laborious task of monitoring widespread deformation, which is critical for ensuring safety. This study utilized the Persistent Scatterer Interferometric Synthetic Aperture Radar (PS-InSAR) technique with high-resolution TerraSAR-X satellite images to measure and analyze deformation at the operating track. Tailored through parameter optimization to fit regional topography and validated against field measurements, PS-InSAR was applied across a 30km by 50km area over two years. The analysis demonstrated that PS-InSAR could efficiently generate high-accuracy time-series data, capable of detecting both the uplift and subsidence processes. It highlighted the importance of augmenting single-point subsidence values with comprehensive time-series analysis for a complete deformation assessment. The study concluded that PS-InSAR is an accurate and cost-effective tool for large-scale linear infrastructure monitoring, despite technological constraints such as radar imaging frequency and the lack of high-resolution sources. Consideringthese constraints, future research will prioritize developing an enhanced algorithm capable of analyzing both urban and suburban areas, accommodating varying numbers of point scatterers. [ABSTRACT FROM AUTHOR]

We develop two models for the temporal evolution of extreme events of multivariate k$$ k $$th order Markov processes. The foundation of our methodology lies in the conditional extremes model of Heffernan and Tawn (Journal of the Royal Statistical Society: Series B (Methodology), 2014, 66, 497–546), and it naturally extends the work of Winter and Tawn (Journal of the Royal Statistical Society: Series C (Applied Statistics), 2016, 65, 345–365; Extremes, 2017, 20, 393–415) and Tendijck et al. (Environmetrics 2019, 30, e2541) to include multivariate random variables. We use cross‐validation‐type techniques to develop a model order selection procedure, and we test our models on two‐dimensional meteorological‐oceanographic data with directional covariates for a location in the northern North Sea. We conclude that the newly‐developed models perform better than the widely used historical matching methodology for these data. [ABSTRACT FROM AUTHOR]

In this paper, we introduce Wide-TSNet, a novel hybrid approach for predicting Bitcoin prices using time-series data transformed into images. The method involves converting time-series data into Markov transition fields (MTFs), enhancing them using histogram equalization, and classifying them using Wide ResNets, a type of convolutional neural network (CNN). We propose a tripartite classification system to accurately represent Bitcoin price trends. In addition, we demonstrate the effectiveness of Wide-TSNet through various experiments, in which it achieves an Accuracy of approximately 94% and an F1 score of 90%. It is also shown that lightweight CNN models, such as SqueezeNet and EfficientNet, can be as effective as complex models under certain conditions. Furthermore, we investigate the efficacy of other image transformation methods, such as Gramian angular fields, in capturing the trends and volatility of Bitcoin prices and revealing patterns that are not visible in the raw data. Moreover, we assess the effect of image resolution on model performance, emphasizing the importance of this factor in image-based time-series classification. Our findings explore the intersection between finance, image processing, and deep learning, providing a robust methodology for financial time-series classification. [ABSTRACT FROM AUTHOR]

This paper proposes a new approach to classify and evaluate defects in concrete structures automatically. To overcome the limitations of defect detection methods that traditionally relied on expert visual observation, the reflection signal of electromagnetic pulses is extracted as time-series data and used to analyze the propagation characteristics of each defect. This study uses deep learning models to analyze these time-series data and classify defects. Since anomaly detection data has more normal data than anomaly data, data augmentation methods such as Time Warping, Noise Injection, Smoothing, Trend Shifting, etc., were applied to solve the problem of data imbalance and overfitting. Among them, Noise Injection showed the best performance. The generalization performance of the proposed method was evaluated through performance evaluation using LSTM, GRU, and TCN models, and LSTM models showed the highest performance. The study results show that the proposed method effectively classifies defect types in concrete structures and can solve the limitations of existing methods by automatic classification through deep learning models. In addition, it was confirmed that the model's performance could be improved by improving the amount and diversity of data by selecting and applying appropriate data augmentation methods. The contribution of the research is to present a new approach that automates the defect detection and classification task of concrete structures and provides high accuracy and efficiency. [ABSTRACT FROM AUTHOR]

Crop classification is indispensable for agricultural monitoring and food security, but early-season mapping has remained challenging. Synthetic aperture radar (SAR), such as RADARSAT Constellation Mission (RCM) and Sentinel-1, can meet higher requirements on the reliability of satellite data acquisition with all-weather and all-day imaging capability to supply dense observations in the early crop season. This study applied the local window attention transformer (LWAT) to time-series SAR data, including RCM and Sentinel-1, for early-season crop classification. The performance of this integration was evaluated over crop-dominated regions (corn, soybean and wheat) in southwest Ontario, Canada. Comparative analyses against several machine learning and deep learning methods revealed the superiority of the LWAT, achieving an impressive F1-score of 97.96% and a Kappa coefficient of 97.08% for the northern crop region and F1-scores of 98.07% and 97.02% for the southern crop region when leveraging time-series data from RCM and Sentinel-1, respectively. Additionally, by the incremental procedure, the evolution of accuracy determined by RCM and Sentinel-1 was analyzed, which demonstrated that RCM performed better at the beginning of the season and could achieve comparable accuracy to that achieved by utilizing both datasets. Moreover, the beginning of stem elongation of corn was identified as a crucial phenological stage to acquire acceptable crop maps in the early season. This study explores the potential of RCM to provide reliable prior information early enough to assist with in-season production forecasting and decision making. [ABSTRACT FROM AUTHOR]

The forest structure of Canada's northern unmanaged boreal forests are not well characterized despite their critical importance for ecosystem services such as carbon storage and caribou habitat, as well as their vulnerability to climate change and increasing disturbance rates, especially wildfire. Spaceborne lidar observations acquired from the Advanced Topographic Laser Altimeter System (ATLAS) instrument onboard the Ice, Cloud and Elevation Satellite-2 (ICESat-2), enables monitoring of forest structure in these remote northern ecosystems and characterization of long-term post-fire recovery; however, ICESat-2 is a sampling instrument, acquiring data transects rather than wall-to-wall data coverage. By imputing ICESat-2 estimates of canopy height using annual time-series (1984–2021) of spectral indices, change metrics, and land cover metrics from Landsat surface reflectance composites, we derived annual, spatially explicit, wall-to-wall estimates of canopy height for the period 1984–2021 across a 19.6 Mha area of boreal forests in northwestern Ontario and a 630,000 ha area of managed boreal forest further south that had coincident airborne laser scanning (ALS) data for validation. The accuracy of the imputation of derived canopy height estimates in the northwestern site was assessed using a reserved set of ICESat-2 observations (r = 0.76; RMSD = 3.04 m, pRMSD = 33.9%, MD = −0.10 m, and pMD = −1.0%). Using coincident ALS data, the accuracy of the imputed canopy heights in the southern study site were also assessed (r = 0.81; RMSD = 3.65 m, pRMSD = 25.87%), MD = −0.70 m, pMD = 4.96%). Examination of height dynamics post fire indicated that canopy height decreased 8–10 years post fire and recovered between 60% and 85% of pre-fire canopy height within 30 years post fire. The approach presented could readily be extended to similar northern boreal forest areas to provide broad-scale characterizations of fire impacts on forest structure and subsequent recovery. [ABSTRACT FROM AUTHOR]

Frameworks for human activity recognition (HAR) can be applied in the clinical environment for monitoring patients' motor and functional abilities either remotely or within a rehabilitation program. Deep Learning (DL) models can be exploited to perform HAR by means of raw data, thus avoiding time-demanding feature engineering operations. Most works targeting HAR with DL-based architectures have tested the workflow performance on data related to a separate execution of the tasks. Hence, a paucity in the literature has been found with regard to frameworks aimed at recognizing continuously executed motor actions. In this article, the authors present the design, development, and testing of a DL-based workflow targeting continuous human activity recognition (CHAR). The model was trained on the data recorded from ten healthy subjects and tested on eight different subjects. Despite the limited sample size, the authors claim the capability of the proposed framework to accurately classify motor actions within a feasible time, thus making it potentially useful in a clinical scenario. [ABSTRACT FROM AUTHOR]

The identification, classification and mapping of different plant communities and habitats is of fundamental importance for defining biodiversity monitoring and conservation strategies. Today, the availability of high temporal, spatial and spectral data from remote sensing platforms provides dense time series over different spectral bands. In the case of supervised mapping, time series based on classical vegetation indices (e.g., NDVI, GNDVI, ...) are usually input characteristics, but the selection of the best index or set of indices (which guarantees the best performance) is still based on human experience and is also influenced by the study area. In this work, several different time series, based on Sentinel-2 images, were created exploring new combinations of bands that extend the classic basic formulas as the normalized difference index. Multivariate Functional Principal Component Analysis (MFPCA) was used to contemporarily decompose the multiple time series. The principal multivariate seasonal spectral variations identified (MFPCA scores) were classified by using a Random Forest (RF) model. The MFPCA and RF classifications were nested into a forward selection strategy to identify the proper and minimum set of indices' (dense) time series that produced the most accurate supervised classification of plant communities and habitat. The results we obtained can be summarized as follows: (i) the selection of the best set of time series is specific to the study area and the habitats involved; (ii) well-known and widely used indices such as the NDVI are not selected as the indices with the best performance; instead, time series based on original indices (in terms of formula or combination of bands) or underused indices (such as those derivable with the visible bands) are selected; (iii) MFPCA efficiently reduces the dimensionality of the data (multiple dense time series) providing ecologically interpretable results representing an important tool for habitat modelling outperforming conventional approaches that consider only discrete time series. [ABSTRACT FROM AUTHOR]

The recent launch of Landsat-9 satellite enriches the opportunities to work with dense time series of multispectral medium-resolution images. The integration of Landsat-9 in a multi-constellation series with Landsat-8 and Sentinel-2 requires a harmonization of the surface reflectance values that can be obtained from the official Level-2 products. This paper proposes the coefficients of the optimal linear transformations for the European continent, which allow to integrate Landsat-9 with the similar operating missions. These coefficients are based on a regression over 30 independent random extractions of 240,000 samples from images of the same areas but acquired by different sensors within two days. The coefficients were validated on an independent dataset. Furthermore, the effects of the proposed harmonization were tested on four popular vegetation indices, by evaluating the distributions of the differences in values obtained from each sensor pair. Finally, a test on a local scale was carried out with a spectroradiometer survey on 16 locations to collect some reference spectra to be compared with the reflectance values provided by the images. The results demonstrate the interoperability of Landsat and Sentinel-2 missions, since reflectance differences are in most cases within the accuracy specifications of the sensors. However, some discrepancies are observed in the blue and SWIR bands, probably due to inconsistencies in the atmospheric correction processes. [ABSTRACT FROM AUTHOR]

Many deep learning works on financial time-series forecasting focus on predicting future prices/returns of individual assets with numerical price-related information for trading, and hence propose models designed for univariate, single-task, and/or unimodal settings. Forecasting for investment and risk management involves multiple tasks in multivariate settings: forecasts of expected returns and risks of assets in portfolios, and correlations between these assets. As different sources/types of time-series influence future returns, risks, and correlations of assets in different ways, it is also important to capture time-series from different modalities. Hence, this article addresses financial time-series forecasting for investment and risk management in a multivariate, multitask, and multimodal setting. Financial time-series forecasting, however, is challenging due to the low signal-to-noise ratios typical in financial time-series, and as intra-series and inter-series relationships of assets evolve across time. To address these challenges, our proposed Temporal Implicit Multimodal Network (TIME) model learns implicit inter-series relationship networks between assets from multimodal financial time-series at multiple time-steps adaptively. TIME then uses dynamic network and temporal encoding modules to jointly capture such evolving relationships, multimodal financial time-series, and temporal representations. Our experiments show that TIME outperforms other state-of-the-art models on multiple forecasting tasks and investment and risk management applications. [ABSTRACT FROM AUTHOR]