Daily Dataset Research Articles

Risk factors-adjusted performance and persistence during the post-subprime crisis period: evidence from Indian mutual fund industry

PurposeThe main objective of the paper is to find evidence of abnormal returns and performance persistence of actively managed equity funds in the Indian context on an annual basis during the post-subprime crisis period between 2009 and 10 and 2019 and 2020.Design/methodology/approachThe study is exploratory and empirical, used daily net asset value (NAV) of 180 equity funds for 10 years and applied the risk-adjusted, Jensen's (1968) single-factor, Fama and French's (1993) three-factor model and Carhart's (1997) four-factor model to evaluate the performance. The performance persistence has been tested using cross-section regression (Bollen and Busse, 2005), the non-parametric contingency approach, along with the robustness measure, i.e. Malkiel's (1995) Z-score, Brown and Goetzmann's (1995) cross-product ratio (CPR) and Kahn and Rudd's (1995) χ2 value.FindingsThe results show that the Indian equity funds are unable to generate abnormal returns, and the size, value and momentum strategies applied by the fund managers in generating abnormal returns do not work effectively. However, funds provide strong evidence of significant performance persistence on an annual basis in the short-term, mid-term and long-term periods. Both parametric as well as non-parametric tests provide identical evidence of persistence, and the performance persistence is independent of the choice of models, as all the models (i.e. two, three or four-factor models) provide significant evidence of persistence.Research limitations/implicationsThough the study is comprehensive and covered a longer period, there is a scope for future research by examining the influence of fund characteristics, fund rating and macroeconomic factors on performance and persistence. It can be extended over to a longer period covering the post-COVID-19 period, a larger sample size and a comparative study of Indian and foreign mutual funds (MFs).Practical implicationsThe outcomes of this research paper can help wealth-maximizing investors in the identification of persistent equity funds and can apply the previous period’s performance information as a useful investment strategy to generate higher returns in the future. We believe that these outcomes will have significant ramifications for all MF stakeholders and policymakers, especially for the Indian industry in ensuring and establishing the credibility of MF managers, in providing better returns as well as to make MF investment more attractive to Indian retail investors.Originality/valueDespite the exponential growth in the Indian MF industry, limited evidence is available on performance and persistence covering a large sample size during the post-sub-prime crisis period using different return models and parametric and non-parametric approaches. The study is based on the daily data set of a larger sample size representing all the Asset Management Company (AMC) and the longer period following the post-subprime crises, which affected capital flows significantly. Moreover, the application of all the measures enables us to understand performance persistence in a larger context.

A defect classification algorithm for gas tungsten arc welding process based on unsupervised learning and few-shot learning strategy

Welding defect prediction is the foundation for ensuring welding quality in gas tungsten arc welding (GTAW). In the prediction process, method based on molten pool vision is the most effective. Since the classification of molten pool defects relies on a substantial volume of labeled data, it is challenging for the models to be applied industrially. This paper presents an algorithm, FS-Classifier, that can achieve high prediction accuracy based on a limited amount of labeled data. The FS-Classifier comprises two stages: Firstly, an unsupervised training approach named RaP is designed to pre-train the feature extractor using extensive unlabeled daily datasets. The RaP consists of a rotation angle prediction task and a position prediction task, which ensure that the network focuses on salient features and precise elements, respectively. Secondly, the support vectors constructed from limited labeled data are used for the feature classifier. The input data is classified to certain class by computing its distances to support vector. The model achieves an accuracy of 94.5 % on the private dataset and 92.8 % on the public dataset for the six classes of defects using 5 % of labeled data volume. In addition, comparative experiments show that our method only requires 5 % of labeled data to achieve accuracy comparable to traditional supervised learning methods. The proposed algorithm addresses the issue of relying on a substantial amount of labeled data in welding process defect classification.

Unveiling asymmetric return spillovers with portfolio implications among Indian stock sectors during Covid-19 pandemic

This paper aims to provide a systematic inquiry into the return spillover dynamics between a network of Indian sectoral indices during the pre- and post-pandemic periods. To analyze the same, this paper uses the asymmetric time-varying parameter vector autoregressions (TVP-VAR) framework. Furthermore, in the spirit of Broadstock et al. (2020), we perform dynamic portfolio exercises based on common hedging techniques and the minimum connectedness portfolio approach to determine what better captures asymmetry. Our daily dataset includes 12 sectoral stocks spanning from January 01, 2017, to May 5, 2023. The findings reveal that negative connectedness dominates throughout the sample period, demonstrating that profit-maximizing agents and risk-averse investors are more likely to react negatively to news. We also show that in the network, the average net transmitters are the banking and other financial service sectors, whereas the net receivers are the information technology, pharmaceutical, and fast-moving consumer goods sectors throughout the period under consideration. Our results show that the minimum connectedness portfolio (MCoP) approach is a very useful method based on Sharpe ratios, as it is either the first or second most profitable among these three competing methods. These results, therefore, yield valuable insights for policymakers and investors

A high-resolution dataset for future compound hot-dry events under climate change

Global climate change is leading to an increase in compound hot-dry events, significantly impacting human habitats. Analysing the causes and effects of these events requires precise data, yet most meteorological data focus on variables rather than extremes, which hinders relevant research. A daily compound hot-dry events (CHDEs) dataset was developed from 1980 to 2100 under various socioeconomic scenarios, using the latest NASA Earth Exchange Global Daily Downscaled Projections (NEX-GDDP-CMIP6) dataset to address this. The dataset has a spatial resolution of 0.25 degrees (approximately 30 kilometres), including three indicators, namely D (the yearly sum of hot-dry extreme days), prI (the intensity of daily precipitation), and tasI (the intensity of daily temperature). To validate the accuracy of the dataset, we compared observational data from China (National Meteorological Information Center, NMIC), Europe (ERA5), and North America (ERA5). Results show close alignment with estimated values from the observational daily dataset, both temporally and spatially. The predictive interval (PI) pass rates for the CHDEs dataset exhibit notably high values. For a 90% PI, D has a pass rate exceeding 85%, whilst prI and tasI respectively show a pass rate above 70% and 95%. These results underscore its suitability for conducting global and regional studies about compound hot-dry events.

Electric load prediction using SARIMAX

ABSTRACT Electricity demand variation is one of the prevalent phenomena in Nepal. This variation can cause excess electricity during low-demand periods and a shortage during high-demand periods. An early prediction of electricity demand allows for levelling such fluctuations by introducing dynamic tariffs based on varying demands. This study aims to use Python’s time series models to make predictions about electrical demands and propose a dynamic tariff system. The dataset of Panama City’s hourly electrical load is brought into use. Two models, namely the Seasonal Auto-Regressive Integrated Moving Average with Exogenous Factors (SARIMAX) and the BATS model, are deemed to be the two best time series models. The selection of optimal seasonal and non-seasonal orders in SARIMAX is made by comparing Bayes Information Criterion (BIC). For the SARIMAX model, a daily dataset has been taken by summing the hourly data for each day, whereas for the BATS model, hourly data for the year 2018 has been used. The SARIMAX model resulted in the best fit with the evaluation metrics values: Root Mean Square Error (RMSE) of 975,336.88, Mean Average error (MAE) of 739.09, and R2 score of 0.75. However, SARIMAX’s limitation is that the use of daily data does not allow dynamic tariff prediction, which is possible with the BATS model.

Stable-isotope variability in daily precipitation: insights from a low-cost collector in SE England

ABSTRACT Precipitation stable-isotope data are often used in hydroclimatic, hydrological and hydrogeological investigations, with measurements typically undertaken on integrated monthly samples. However, daily sampling reveals overlooked aspects of controls on precipitation isotope values, including synoptic meteorological conditions. We present a one-year record of stable isotopes in daily precipitation during 2021, from a site in SE England close to Greater London. We find marked daily variability over the course of the year (–15.62 to +0.92 ‰ for δ18O, –108.7 to +2.9 ‰ for δ2H and –6.5 to +23.1 ‰ for deuterium excess). Correlations with individual meteorological variables including precipitation amount, temperature and weather type are moderate to weak suggesting complex controls on the daily rainfall isotope values. The daily data are compared with three other daily datasets from England and, by conversion to monthly values, directly with data from three long-term collection stations across Britain and Ireland. The scale of variability in the daily data from our site is consistent with that seen in other English records despite them all coming from different time periods. The monthly data show broad consistency, although there are differences that also highlight geographical variability in precipitation values across the British Isles.

Examining market synchronicity: spillover connectedness among commodity markets and exchange rates during crisis periods

We investigate the dynamic spillover connectedness between commodity markets (gold and crude oil) and exchange rates in seven of the world's most influential developed and emerging economies: China, Ukraine, Germany, the United Kingdom, Japan, Russia, and the United States. Utilizing an extensive daily dataset sourced from Bloomberg, we analyze key exchange rates and commodity markets, focusing on currencies and economic indicators of significant global economies. Our findings reveal significant spillover effects across markets, with major currencies playing central roles in transmitting volatility during crisis periods. Specifically, currencies like EUR/$ and GBP/$ exhibit substantial internal spillovers, reflecting robust interconnectedness within these markets. In contrast, commodities such as crude oil demonstrate a more insulated nature, transmitting fewer spillovers compared to exchange rates. Furthermore, our analysis during distinct crisis periods, including the Covid-19 pandemic, the Russia-Ukraine conflict, and the Israel-Palestine war, highlights evolving market dynamics amidst geopolitical uncertainties. Major currencies continue to play pivotal roles during crises, transmitting the highest total spillovers, while commodities like gold emerge as significant transmitters and receivers, underscoring their role as safe-haven assets.

High resolution mapping of nitrogen dioxide and particulate matter in Great Britain (2003–2021) with multi-stage data reconstruction and ensemble machine learning methods

In this contribution, we applied a multi-stage machine learning (ML) framework to map daily values of nitrogen dioxide (NO2) and particulate matter (PM10 and PM2.5) at a 1 km2 resolution over Great Britain for the period 2003–2021. The process combined ground monitoring observations, satellite-derived products, climate reanalyses and chemical transport model datasets, and traffic and land-use data. Each feature was harmonized to 1 km resolution and extracted at monitoring sites. Models used single and ensemble-based algorithms featuring random forests (RF), extreme gradient boosting (XGB), light gradient boosting machine (LGBM), as well as lasso and ridge regression. The various stages focused on augmenting PM2.5 using co-occurring PM10 values, gap-filling aerosol optical depth and columnar NO2 data obtained from satellite instruments, and finally the training of an ensemble model and the prediction of daily values across the whole geographical domain (2003–2021). Results show a good ensemble model performance, calculated through a ten-fold monitor-based cross-validation procedure, with an average R2 of 0.690 (range 0.611–0.792) for NO2, 0.704 (0.609–0.786) for PM10, and 0.802 (0.746–0.888) for PM2.5. Reconstructed pollution levels decreased markedly within the study period, with a stronger reduction in the latter eight years. The pollutants exhibited different spatial patterns, while NO2 rose in close proximity to high-traffic areas, PM demonstrated variation at a larger scale. The resulting 1 km2 spatially resolved daily datasets allow for linkage with health data across Great Britain over nearly two decades, thus contributing to extensive, extended, and detailed research on the long-and short-term health effects of air pollution.

Spillover effect of the geopolitical uncertainty on the cryptocurrency market

PurposeThis study aims to examine the influence of geopolitical uncertainty on cryptocurrency markets (CM).Design/methodology/approachUtilizing two distinct sets of daily returns data spanning from January 1, 2019, to May 4, 2023, the analysis employs the geopolitical risk (GPR) index formulated by Caldara and Iacoviello (2022), which encapsulates two pivotal events: the COVID-19 pandemic and the Russia–Ukraine conflict. The cryptocurrency market (CM) encompasses Bitcoin (BTC), Ethereum (ETH), Litecoin (LTC) and Dogecoin (DOGE). Employing the DCC-GARCH model and supplementing it with wavelet coherence analysis to discern perceptual distinctions between short- and long-term market reactions.FindingsThe main findings indicate that the GPR index clearly impacts the return of CM in the short-, mid- and long-term periods. BTC exhibited the highest volatility in response to changes in the GPR index. The cryptocurrency market offers a better diversification opportunity, and the impact of geopolitical events varies across time, with their direction and magnitude closely related to the specificity of the CM.Practical implicationsThis research is helpful for financial market investors, portfolio and risk managers, make informed decisions about including cryptocurrencies in their investment portfolios to mitigate the risks in uncertainty period.Originality/valueCryptocurrency market volatility is treated weakly during the risk period. With advanced statistical method, this study links two important events: the COVID-19 pandemic and the Russia–Ukraine conflict and selects the top four cryptocurrencies constituting 80% of the market. This study examines the impact of geopolitical risk on the cryptocurrency market and shows that this market is considered a safe haven.

Hedging Carbon Price Risk on EU ETS: A Comparison of Green Bonds from the EU, US, and China

This article aims to examine the hedging effect of green bonds in the US market, the European market, and the Chinese market on carbon price risk in the European Union Emission Trading System (EU ETS) from 2021 to 2023. By using daily datasets extracted from Bloomberg and the Vector Error Correction Model (VECM), the research provides evidence of the hedging effect of green bonds in all three markets on carbon price risk in the EU ETS. The paper concludes that the hedging ratio is positive for green bonds in the EU and China, while the figure for the US market is negative. Moreover, there is a positive effect of oil prices on carbon returns in EU ETS. Meanwhile, the opposite is found for stock prices.

How do seasonal, significant events, and policies affect China's REE export prices? Based on deep learning perspective

This paper aims to accurately predict China's rare earth export prices and reveal the impact of variables such as seasonality, significant events, finance, and supply and demand on rare earth price volatility. Daily datasets of light and heavy rare earths from 2011 to 2023 were used, and the Tree-structured Parzen Estimator-Temporal Fusion Transformer model was employed to predict rare earth prices. Complete Ensemble Empirical Mode Decomposition with Adaptive Noise and partial dependence plots were used to reveal the factors affecting price volatility. The following conclusions were drawn: (1) The Tree-structured Parzen Estimator-Temporal Fusion Transformer deep learning model can provide more accurate rare earth price prediction information; (2) Light rare earth prices are more susceptible to cyclical influences, while heavy rare earth prices are more affected by significant events. The outbreak of COVID-19 has had a long-term impact on both light and heavy rare earth prices; (3) The fluctuations in heavy rare earth prices are mainly influenced by financial factors, while the fluctuations in light rare earth prices are influenced by multiple factors such as finance, supply and demand, and macroeconomics; (4) An increase in resource tax rates may lead to a decrease in rare earth prices, while an increase in restrictions on rare earth mining may lead to an increase in rare earth prices.

COVID-19 pandemic and linkage between stock markets in Middle Eastern countries

PurposeThe purpose of this study is to evaluate the linkage between stock markets in Middle Eastern countries before and during the COVID-19 pandemic by using daily and monthly data sets for the period from 2011 to 2021.Design/methodology/approachThe multivariate BEKK-GARCH model was computed to evaluate the existence of non-linear linkage among Middle Eastern stock markets. A correlation approach was used in this study to determine the type of linear connectivity between Middle Eastern stock markets. The study used monthly and daily data sets covering the years 2011 to 2021 to investigate the linkage between stock returns and the volatility spillover between the stock markets in Palestine, Jordan, Syria and Lebanon, both before and during COVID-19. To understand the types of relationships between markets before and during COVID-19, the daily data set was split into two periods.FindingsResults from the pre-COVID-19 suggest that the Syria stock market is not related to any stock market in the Middle East markets; the Palestine and Lebanon stock markets exhibit a weak relationship, but Jordan and Palestine stock markets are strongly linked. Conversely, results from COVID-19 evince a very strong bidirectional volatility spillover between Middle East stock markets. Overall, the results indicate the existence of increased linkage during the COVID-19.Research limitations/implicationsThe data collection on a daily and monthly basis, both before and during COVID-19, presents certain limitations for the paper. Another limitation is that the data cannot be generalized to all other Middle Eastern countries; rather, the conclusions drawn can only be applied to these four countries. This is especially true if the scholars collected most of the necessary data but were unable to obtain certain data for various reasons.Practical implicationsThese findings have implications for risk management, market regulation and the growth of local stock markets. Facilitating the growth of smaller, more specialized markets to improve integration with other Middle Eastern markets is one of the goals of the domestic stock market development policy. To ensure financial stability, Middle Eastern stock market linking policies should consider spillover risk and take steps to minimize it. Enhancing the range of investment opportunities accessible to shareholders and functioning as confidential risk-sharing mechanisms to facilitate improved risk management in Middle Eastern stock markets will not only significantly influence the mobilization of private capital to promote investment and local economic growth but also lay groundwork for integrated market platforms.Originality/valueThis paper adds to the body of literature by demonstrating the nature of the connections between these small markets and the larger markets in the Middle East region. Information from the smaller markets provides institutional insights that enhance the body of existing research, guide the formulation of evidence-based policies and advance financial literacy in these markets. This study contributes by comparing data from different stock markets to better understand the type and strength of the link and relationship between Middle Eastern stock markets, as well as any underlying or reinforcing factors that might have contributed to the relationship and the specific types of links that these markets shared prior and during COVID-19.

IPB-MSA&SO4: a daily 0.25° resolution dataset of in situ-produced biogenic methanesulfonic acid and sulfate over the North Atlantic during 1998–2022 based on machine learning

Abstract. Accurate long-term marine-derived biogenic sulfur aerosol concentrations at high spatial and temporal resolutions are critical for a wide range of studies, including climatology, trend analysis, and model evaluation; this information is also imperative for the accurate investigation of the contribution of marine-derived biogenic sulfur aerosol concentrations to the aerosol burden, for the elucidation of their radiative impacts, and to provide boundary conditions for regional models. By applying machine learning algorithms, we constructed the first publicly available daily gridded dataset of in situ-produced biogenic methanesulfonic acid (MSA) and non-sea-salt sulfate (nss-SO4=) concentrations covering the North Atlantic. The dataset is of high spatial resolution (0.25° × 0.25°) and spans 25 years (1998–2022), far exceeding what observations alone could achieve both spatially and temporally. The machine learning models were generated by combining in situ observations of sulfur aerosol data from Mace Head Atmospheric Research Station, located on the west coast of Ireland, and from the North Atlantic Aerosols and Marine Ecosystems Study (NAAMES) cruises in the northwestern Atlantic with the constructed sea-to-air dimethylsulfide flux (FDMS) and ECMWF ERA5 reanalysis datasets. To determine the optimal method for regression, we employed five machine learning model types: support vector machines, decision tree, regression ensemble, Gaussian process regression, and artificial neural networks. A comparison of the mean absolute error (MAE), root-mean-square error (RMSE), and coefficient of determination (R2) revealed that Gaussian process regression (GPR) was the most effective algorithm, outperforming the other models with respect to simulating the biogenic MSA and nss-SO4= concentrations. For predicting daily MSA (nss-SO4=), GPR displayed the highest R2 value of 0.86 (0.72) and the lowest MAE of 0.014 (0.10) µg m−3. GPR partial dependence analysis suggests that the relationships between predictors and MSA and nss-SO4= concentrations are complex rather than linear. Using the GPR algorithm, we produced a high-resolution daily dataset of in situ-produced biogenic MSA and nss-SO4= sea-level concentrations over the North Atlantic, which we named “In-situ Produced Biogenic Methanesulfonic Acid and Sulfate over the North Atlantic” (IPB-MSA&SO4). The obtained IPB-MSA&SO4 data allowed us to analyze the spatiotemporal patterns of MSA and nss-SO4= as well as the ratio between them (MSA:nss-SO4=). A comparison with the existing Copernicus Atmosphere Monitoring Service ECMWF Atmospheric Composition Reanalysis 4 (CAMS-EAC4) reanalysis suggested that our high-resolution dataset reproduces the spatial and temporal patterns of the biogenic sulfur aerosol concentration with high accuracy and has high consistency with independent measurements in the Atlantic Ocean. IPB-MSA&SO4 is publicly available at https://doi.org/10.17632/j8bzd5dvpx.1 (Mansour et al., 2023b).

Classifying Self-Reported Rheumatoid Arthritis Flares Using Daily Patient-Generated Data From a Smartphone App: Exploratory Analysis Applying Machine Learning Approaches.

The ability to predict rheumatoid arthritis (RA) flares between clinic visits based on real-time, longitudinal patient-generated data could potentially allow for timely interventions to avoid disease worsening. This exploratory study aims to investigate the feasibility of using machine learning methods to classify self-reported RA flares based on a small data set of daily symptom data collected on a smartphone app. Daily symptoms and weekly flares reported on the Remote Monitoring of Rheumatoid Arthritis (REMORA) smartphone app from 20 patients with RA over 3 months were used. Predictors were several summary features of the daily symptom scores (eg, pain and fatigue) collected in the week leading up to the flare question. We fitted 3 binary classifiers: logistic regression with and without elastic net regularization, a random forest, and naive Bayes. Performance was evaluated according to the area under the curve (AUC) of the receiver operating characteristic curve. For the best-performing model, we considered sensitivity and specificity for different thresholds in order to illustrate different ways in which the predictive model could behave in a clinical setting. The data comprised an average of 60.6 daily reports and 10.5 weekly reports per participant. Participants reported a median of 2 (IQR 0.75-4.25) flares each over a median follow-up time of 81 (IQR 79-82) days. AUCs were broadly similar between models, but logistic regression with elastic net regularization had the highest AUC of 0.82. At a cutoff requiring specificity to be 0.80, the corresponding sensitivity to detect flares was 0.60 for this model. The positive predictive value (PPV) in this population was 53%, and the negative predictive value (NPV) was 85%. Given the prevalence of flares, the best PPV achieved meant only around 2 of every 3 positive predictions were correct (PPV 0.65). By prioritizing a higher NPV, the model correctly predicted over 9 in every 10 non-flare weeks, but the accuracy of predicted flares fell to only 1 in 2 being correct (NPV and PPV of 0.92 and 0.51, respectively). Predicting self-reported flares based on daily symptom scorings in the preceding week using machine learning methods was feasible. The observed predictive accuracy might improve as we obtain more data, and these exploratory results need to be validated in an external cohort. In the future, analysis of frequently collected patient-generated data may allow us to predict flares before they unfold, opening opportunities for just-in-time adaptative interventions. Depending on the nature and implication of an intervention, different cutoff values for an intervention decision need to be considered, as well as the level of predictive certainty required.

Gauging the dynamic interlinkage level between Fintech and the global blue economy performance

PurposeWe investigate connections between the development of Fintech and the blue economy from September 14th, 2020, to August 11th, 2023.Design/methodology/approachIn this research, we use a cutting-edge model-free connectedness approach to investigate the relationships between FinTech and blue bond volatility. Our work is the first to investigate the effects of unknown events, such as the COVID-19 pandemic and Ukraine–Russia conflicts, on the interconnection of volatility derived from FinTech development and blue bond volatility.FindingsOur results highlight the two-way relationship between the development of Fintech and the blue economy during our sample period. The net total connectedness shows that the blue economy index is a net shock receiver, especially in late 2021 and the second half of 2022, while most of the fintech indexes in our sample are mainly net shock transmitters. The Ukraine–Russia tension threatens the development of a sustainable blue economy. The development of Fintech plays an important role in promoting the blue economy.Practical implicationsOur results have important policy implications for investors and governments, as well as methods from the spillovers across the various indicators and their interconnections. Sharp information on the primary contagions among these indicators aids politicians in designing the most appropriate policies.Originality/valueOur paper contributes to the literature in at least four ways. First, as previously stated, our article is the first to investigate the relationship between FinTech and blue bond volatility. Second, this study presented a framework for studying volatility interconnections between distinct variables that is more suited to analyzing these interconnections. In this research, we use a cutting-edge model-free connectedness approach to investigate the relationships between FinTech and blue bond volatility. Third, our work is the first to investigate the effects of unknown events such as the COVID-19 pandemic and Ukraine–Russia conflicts on the interconnection of volatility deriving from FinTech development and blue bond volatility. Lastly, our research provides a daily dataset for the BNP Paribas Easy ECPI Global ESG Blue Economy UCITS ETF to analyze 50 businesses from various markets that are at the forefront of the responsible application of ocean resources and other ESG standards. The Global X FinTech ETF (FINX) and the ARK FinTech Innovation ETF (ARKF) seek exposure to companies developing financial technology innovations. The development sectors include insurance, investment, fundraising and third-party lending by utilizing cutting-edge mobile and digital technologies. Our time series runs from September 14th, 2020, to August 11th, 2023. By using this database, we provide a comprehensive analysis of the link between the volatilities arising from various markets.

Generative Adversarial Network for Real‐Time Flash Drought Monitoring: A Deep Learning Study

AbstractDroughts are among the most devastating natural hazards, occurring in all regions with different climate conditions. The impacts of droughts result in significant damages annually around the world. While drought is generally described as a slow‐developing hazardous event, a rapidly developing type of drought, the so‐called flash drought has been revealed by recent studies. The rapid onset and strong intensity of flash droughts require accurate real‐time monitoring. Addressing this issue, a Generative Adversarial Network (GAN) is developed in this study to monitor flash droughts over the Contiguous United States (CONUS). GAN contains two models: (a) discriminator and (b) generator. The developed architecture in this study employs a Markovian discriminator, which emphasizes the spatial dependencies, with a modified U‐Net generator, tuned for optimal performance. To determine the best loss function for the generator, four different networks are developed with different loss functions, including Mean Absolute Error (MAE), adversarial loss, a combination of adversarial loss with Mean Square Error (MSE), and a combination of adversarial loss with MAE. Utilizing daily datasets collected from NLDAS‐2 and Standardized Soil Moisture Index (SSI) maps, the network is trained for real‐time daily SSI monitoring. Comparative assessments reveal the proposed GAN's superior ability to replicate SSI values over U‐Net and Naïve models. Evaluation metrics further underscore that the developed GAN successfully identifies both fine‐ and coarse‐scale spatial drought patterns and abrupt changes in the SSI temporal patterns that is important for flash drought identification.

Analysis of Beijing Traffic Violations Based on the BERT-CRF Model

Traffic violations are a major cause of traffic accidents, yet current research falls short in comprehensively analysing these violations and the named entity method fails to extract the name of traffic violation events from records, thereby lacking in providing guidance for managing urban traffic violations. By expanding the People’s Daily dataset from 71,456 words to 95,291 words, the BERT-CRF (Bidirectional Encoder Representations from Transformers-Conditional Random Field) model achieves an accuracy rate of 88.53%, a recall rate of 92.90% and an F1 score of 90.66%, successfully identifying event, time and location named entities within traffic violations. The data of traffic violations is then enhanced through forward geocoding and the Bayesian formula, and traffic violations are analysed from time, space, administrative region, gender and weather, to provide support for the dynamic allocation of law enforcement forces on traffic scenes and the precise management oftraffic violations.

Advanced forecasting of COVID-19 epidemic: Leveraging ensemble models, advanced optimization, and decomposition techniques

In the global effort to address the outbreak of the new coronavirus pneumonia (COVID-19) pandemic, accurate forecasting of epidemic patterns has become crucial for implementing successful interventions aimed at preventing and controlling the spread of the disease. The correct prediction of the course of COVID-19 outbreaks is a complex and challenging task, mainly because of the significant volatility in the data series related to COVID-19. Previous studies have been limited by the exclusive use of individual forecasting techniques in epidemic modeling, disregarding the integration of diverse prediction procedures. The lack of attention to detail in this situation can yield worse-than-ideal results. Consequently, this study introduces a novel ensemble framework that integrates three machine learning methods (kernel ridge regression (KRidge), Deep random vector functional link (dRVFL), and ridge regression) within a linear relationship (L-KRidge-dRVFL-Ridge). The optimization of this framework is accomplished through a distinctive approach, specifically adaptive differential evolution and particle swarm optimization (A-DEPSO). Moreover, an effective decomposition method, known as time-varying filter empirical mode decomposition (TVF-EMD), is employed to decompose the input variables. A feature selection technique, specifically using the light gradient boosting machine (LGBM), is also implemented to extract the most influential input variables. The daily datasets of COVID-19 collected from two countries, namely Italy and Poland, were used as the experimental examples. Additionally, all models are implemented to forecast COVID-19 at two-time horizons: 10- and 14-day ahead (t+10 and t+14). According to the results, the proposed model can yield higher correlation coefficient (R) for both case studies: Italy (t+10 = 0.965, t+14 = 0.961) and Poland (t+10 = 0.952, t+14 = 0.940) than the other models. The experimental results demonstrate that the model suggested in this paper has outstanding results in various kinds of complex epidemic prediction situations. The proposed ensemble model demonstrates exceptional accuracy and resilience, outperforming all similar models in terms of efficacy.

Holistic-Guided Disentangled Learning With Cross-Video Semantics Mining for Concurrent First-Person and Third-Person Activity Recognition.

The popularity of wearable devices has increased the demands for the research on first-person activity recognition. However, most of the current first-person activity datasets are built based on the assumption that only the human-object interaction (HOI) activities, performed by the camera-wearer, are captured in the field of view. Since humans live in complicated scenarios, in addition to the first-person activities, it is likely that third-person activities performed by other people also appear. Analyzing and recognizing these two types of activities simultaneously occurring in a scene is important for the camera-wearer to understand the surrounding environments. To facilitate the research on concurrent first- and third-person activity recognition (CFT-AR), we first created a new activity dataset, namely PolyU concurrent first- and third-person (CFT) Daily, which exhibits distinct properties and challenges, compared with previous activity datasets. Since temporal asynchronism and appearance gap usually exist between the first- and third-person activities, it is crucial to learn robust representations from all the activity-related spatio-temporal positions. Thus, we explore both holistic scene-level and local instance-level (person-level) features to provide comprehensive and discriminative patterns for recognizing both first- and third-person activities. On the one hand, the holistic scene-level features are extracted by a 3-D convolutional neural network, which is trained to mine shared and sample-unique semantics between video pairs, via two well-designed attention-based modules and a self-knowledge distillation (SKD) strategy. On the other hand, we further leverage the extracted holistic features to guide the learning of instance-level features in a disentangled fashion, which aims to discover both spatially conspicuous patterns and temporally varied, yet critical, cues. Experimental results on the PolyU CFT Daily dataset validate that our method achieves the state-of-the-art performance.

Comparison and evaluation of the performance of reanalysis datasets for compound extreme temperature and precipitation events in the Qilian Mountains

Due to potential powerful destruction, the compound extreme temperature and precipitation events have continuously captured the interest of scientists. However, the lack of in-situ datasets limits the knowledge of compound extreme events in the alpine mountains. We selected the Qilian Mountains and analyzed the change trends of four types of compound extreme weather events, and further evaluated the performances of five reanalysis datasets for compound extreme weather events based on the observed daily datasets from 33 meteorological stations from 1980 to 2020. The results showed that (1) The frequency of four compound extreme weather events from highest to lowest is cold-dry events, high temperature-dry events, cold-heavy precipitation events, and high temperature-heavy precipitation events. High temperature-heavy precipitation events and high temperature-dry events showed increasing trends, while cold-dry events and cold-heavy precipitation events showed decreasing trends. (2) All five reanalysis datasets underestimated temperature and overestimated precipitation. In terms of extreme temperature and extreme precipitation events, ERA5-Land performed best, followed by ERA5, Har v2 and Merra2, and JRA55 performed worst. (3) For the four types of compound weather events, the reanalysis datasets performed better for high-temperature or dry events than cold or heavy precipitation events in the Qilian Mountains. The reanalysis datasets performed best for compound high temperature-dry events and worst for compound extreme temperature-heavy precipitation events. According to the results, we recommend ERA5-Land and ERA5 for the research on the compound extreme events. The research results provide theoretical support for the study of compound extreme temperature-precipitation events in alpine regions with sparse in-situ datasets.