The impact of COVID-19 restrictions on HIV prevention and treatment services for key populations in South Africa: an interrupted time series analysis.

BACKGROUND: Key populations (KP), including men who have sex with men (MSM), female sex workers (FSW), and transgender women (TGW), experience a disproportionate burden of HIV, even in generalized epidemics like South Africa. Given this disproportionate burden and unique barriers to accessing health services, sustained provision of care is particularly relevant. It is unclear how the COVID-19 pandemic and its associated restrictions may have impacted this delivery. In this study, we aimed to describe patterns of engagement in HIV prevention and treatment services among KP in South Africa and assess the impact of different COVID-19 restriction levels on service delivery. METHODS: We leveraged programmatic data collected by the US President's Emergency Plan for AIDS Relief (PEPFAR)-supported KP partners in South Africa. We divided data into three discrete time periods based on national COVID-19 restriction periods: (i) Pre-restriction period, (ii) High-level restriction period, and (iii) After-high level restriction period. Primary outcomes included monthly total HIV tests, new HIV cases identified, new initiations of pre-exposure prophylaxis (PrEP), and new enrollments in antiretroviral therapy (ART). We conducted interrupted time series segmented regression analyses to estimate the impact of COVID-19 restrictions on HIV prevention and treatment service utilization. RESULTS: Between January 2018 and June 2022, there were a total of 231,086 HIV tests, 27,051 HIV positive cases, 27,656 pre-exposure prophylaxis (PrEP) initiations, and 15,949 antiretroviral therapy initiations among MSM, FSW and TGW in PEPFAR-supported KP programs in South Africa. We recorded 90,457 total HIV tests during the 'pre-restriction' period, with 13,593 confirmed new HIV diagnoses; 26,134 total HIV tests with 2,771 new diagnoses during the 'high-level restriction' period; and 114,495 HIV tests with 10,687 new diagnoses during the after high-level restriction period. Our Poisson regression model estimates indicate an immediate and significant decrease in service engagement at the onset of COVID-19 restrictions, including declines in HIV testing, treatment, and PrEP use, which persisted. As programs adjusted to the new restrictions, there was a gradual rebound in service engagement, particularly among MSM and FSW. Towards the end of the high-level restriction period, with some aspects of daily life returning to normal but others still restricted, there was more variability. Some indicators continued to improve, while others stagnated or decreased. CONCLUSION: Service provision rebounded from the initial shock created by pandemic-related restrictions, and HIV services were largely maintained for KP in South Africa. These results suggest that HIV service delivery among programs designed for KP was able to be flexible and resilient to the evolving restrictions. The results of this study can inform plans for future pandemics and large-scale disruptions to the delivery of HIV services.

Nursing of forearm hematoma after transradial coronary intervention.

Background: Complications such as forearm hematoma after coronary intervention through the radial artery are a common complication. Material and methods: By observing, describing, and analyzing the pictures taken during clinical diagnosis and consultation, we summarize the prevention, treatment, and nursing of forearm hematoma after percutaneous coronary intervention, to provide reference for the nursing of patients with forearm hematoma. Results: We have innovatively summarized the risk classification of forearm hematoma and the three key time points for preventing hematoma. Conclusion: Complications such as forearm hematoma after coronary intervention through the radial artery are a common complication. We have innovatively summarized the risk classification of forearm hematoma and the three key time points for preventing hematoma, providing reference for the prevention and management of forearm hematoma in clinical practice. For patients undergoing transradial coronary intervention, the three key time points for preventing hematoma and symptomatic management based the risk classification of forearm hematoma are crucial.

Reliable generation of privacy-preserving synthetic electronic health record time series via diffusion models.

OBJECTIVE: Electronic health records (EHRs) are rich sources of patient-level data, offering valuable resources for medical data analysis. However, privacy concerns often restrict access to EHRs, hindering downstream analysis. Current EHR deidentification methods are flawed and can lead to potential privacy leakage. Additionally, existing publicly available EHR databases are limited, preventing the advancement of medical research using EHR. This study aims to overcome these challenges by generating realistic and privacy-preserving synthetic EHRs time series efficiently. MATERIALS AND METHODS: We introduce a new method for generating diverse and realistic synthetic EHR time series data using denoizing diffusion probabilistic models. We conducted experiments on 6 databases: Medical Information Mart for Intensive Care III and IV, the eICU Collaborative Research Database (eICU), and non-EHR datasets on Stocks and Energy. We compared our proposed method with 8 existing methods. RESULTS: Our results demonstrate that our approach significantly outperforms all existing methods in terms of data fidelity while requiring less training effort. Additionally, data generated by our method yield a lower discriminative accuracy compared to other baseline methods, indicating the proposed method can generate data with less privacy risk. DISCUSSION: The proposed model utilizes a mixed diffusion process to generate realistic synthetic EHR samples that protect patient privacy. This method could be useful in tackling data availability issues in the field of healthcare by reducing barrier to EHR access and supporting research in machine learning for health. CONCLUSION: The proposed diffusion model-based method can reliably and efficiently generate synthetic EHR time series, which facilitates the downstream medical data analysis. Our numerical results show the superiority of the proposed method over all other existing methods.

Peritoneal baclofen pump migration: a rare complication of subfascial placement. Illustrative cases.

BACKGROUND: The intrathecal baclofen pump is an effective treatment option for patients with severe spasticity. In children, subfascial pump placement is often preferred given concerns for infection and wound healing. However, this approach is not without risk, and rare complications, such as peritoneal pump migration, can occur. OBSERVATIONS: The authors describe three pediatric cases of peritoneal pump migration at their institution over the past 14 years (3/545, 0.5%). All three patients had low body weight (below the 39th percentile), and two had scoliosis requiring surgery. All pumps had been placed using the subfascial technique. The first case occurred 6 months postplacement, and the pump was not replaced. Cases 2 and 3 occurred at 2 and 3 years postplacement, respectively, and both pumps were replaced. LESSONS: The authors conclude that peritoneal pump migration, although uncommon, can occur in patients with subfascial pump placement, and providers should have a low threshold of suspicion for repeat imaging prior to refilling if the pump's location has migrated. Potential contributing factors to pump migration include a patient's small size, a larger pump size (40-mL pump), and scoliosis. All these factors should be considered during pump placement, and surgeons can consider using a 20-ml pump for smaller patients. https://thejns.org/doi/abs/10.3171/CASE24290.

Multi-dimensional epidemiology and informatics data on COVID-19 wave at the end of zero COVID policy in China.

Background: China exited strict Zero-COVID policy with a surge in Omicron variant infections in December 2022. Given China's pandemic policy and population immunity, employing Baidu Index (BDI) to analyze the evolving disease landscape and estimate the nationwide pneumonia hospitalizations in the post Zero COVID period, validated by hospital data, holds informative potential for future outbreaks. Methods: Retrospective observational analyses were conducted at the conclusion of the Zero-COVID policy, integrating internet search data alongside offline records. Methodologies employed were multidimensional, encompassing lagged Spearman correlation analysis, growth rate assessments, independent sample T-tests, Granger causality examinations, and Bayesian structural time series (BSTS) models for comprehensive data scrutiny. Results: Various diseases exhibited a notable upsurge in the BDI after the policy change, consistent with the broader trajectory of the COVID-19 pandemic. Robust connections emerged between COVID-19 and diverse health conditions, predominantly impacting the respiratory, circulatory, ophthalmological, and neurological domains. Notably, 34 diseases displayed a relatively high correlation (r > 0.5) with COVID-19. Among these, 12 exhibited a growth rate exceeding 50% post-policy transition, with myocarditis escalating by 1,708% and pneumonia by 1,332%. In these 34 diseases, causal relationships have been confirmed for 23 of them, while 28 garnered validation from hospital-based evidence. Notably, 19 diseases obtained concurrent validation from both Granger causality and hospital-based data. Finally, the BSTS models approximated approximately 4,332,655 inpatients diagnosed with pneumonia nationwide during the 2 months subsequent to the policy relaxation. Conclusion: This investigation elucidated substantial associations between COVID-19 and respiratory, circulatory, ophthalmological, and neurological disorders. The outcomes from comprehensive multi-dimensional cross-over studies notably augmented the robustness of our comprehension of COVID-19's disease spectrum, advocating for the prospective utility of internet-derived data. Our research highlights the potential of Internet behavior in predicting pandemic-related syndromes, emphasizing its importance for public health strategies, resource allocation, and preparedness for future outbreaks.

How much rainwater contributes to a spring discharge in the Guarani Aquifer System: insights from stable isotopes and a mass balance model.

Outcrops play an important role in groundwater recharge. Understanding groundwater origins, dynamics and its correlation with different water sources is essential for effective water resources management and planning in terms of quantity and quality. In the case of the Guarani Aquifer System (GAS) outcrop areas are particularly vulnerable to groundwater pollution due to direct recharge processes. This study focuses on the Alto Jacaré-Pepira sub-basin, a watershed near Brotas, a city in the central region of the state of São Paulo, Brazil, where groundwater is vital for supporting tourism, agriculture, urban water supply, creeks, river and wetlands. The area has a humid tropical climate with periods of both intense rainfall and drought, and the rivers remain perennial throughout the year. Therefore, the aim of this study is to investigate the interconnections between a spring and its potential sources of contribution, namely rain and groundwater, in order to elucidate the relationships between the different water sources. To achieve this, on-site monitoring of groundwater depth, rainfall amount, and stable isotope ratios (deuterium (2H) and oxygen-18 (18O)) from rain, spring discharge, and a monitoring well was carried out from 2013 to 2021. The results indicate that the mean and standard deviations for δ18O in rainwater exhibit higher variability, resulting in -4.49 ± 3.18 ‰ VSMOW, while δ18O values from the well show minor variations, similar to those of the spring, recording -7.25 ± 0.32 ‰ and -6.94 ± 0.28 ‰ VSMOW, respectively. The mixing model's outcomes reveal seasonal variations in water sources contribution and indicate that groundwater accounts for approximately 80 % of spring discharge throughout the year. Incorporating stable isotopes into hydrological monitoring provides valuable data for complementing watershed analysis. The values obtained support the significance of the aquifer as a primary source, thereby offering critical insights into stream dynamics of the region.

Temporal patterns in coral disease prevalences at Thudufushi Island, Maldives, 2010-2022.

Coral reefs are lately suffering a fast decline in biodiversity due to the coupled effect of climate change and disease outbreaks, which in recent decades have been reported with higher frequency and shorter intervals. Limited studies have been conducted on coral diseases in the Maldives resulting in the impossibility of assessing the temporal trend in their dynamics. In this context, we evaluated the change in the distribution, prevalence, and host range of 4 diseases, namely black band disease (BBD), brown band disease (BrB), skeletal eroding band (SEB) and white syndrome (WS), in the reef system around Thudufushi Island after an interval of 12 yr since the last assessment. In this period, the overall disease prevalence increased, except for BrB, with SEB showing the most severe increase in 2022 in comparison to 2010. The overall average prevalence of coral diseases is approximately 2%, indicating an increase of about 0.7% since 2010. Diseased coral colonies were found in all the investigated sites, with the east site being the most affected and SEB emerging as the most prevalent disease across all the investigated sites. The affected colonies belong to 13 genera, with Psammocora genus showing the highest overall mean disease prevalence. This study depicted a basic temporal trend in disease prevalence that confirms an increase in coral diseases in the region and calls for a dedicated national monitoring protocol to better understand and predict future coral disease dynamics at regional scales.

Ambient air pollution contributed to pulmonary tuberculosis in China.

Published studies on outdoor air pollution and tuberculosis risk have shown heterogeneous results. Discrepancies in prior studies may be partially explained by the limited geographic scope, diverse exposure times, and heterogeneous statistical methods. Thus, we conducted a multi-province, multi-city time-series study to comprehensively investigate this issue. We selected 67 districts or counties from all geographic regions of China as study sites. We extracted data on newly diagnosed pulmonary tuberculosis (PTB) cases, outdoor air pollutant concentrations, and meteorological factors in 67 sites from January 1, 2014 to December 31, 2019. We utilized a generalized additive model to evaluate the relationship between ambient air pollutants and PTB risk. Between 2014 and 2019, there were 172 160 newly diagnosed PTB cases reported in 67 sites. With every 10-µg/m3 increase in SO2, NO2, PM10, PM2.5, and 1-mg/m3 in CO, the PTB risk increased by 1.97% [lag 0 week, 95% confidence interval (CI): 1.26, 2.68], 1.30% (lag 0 week, 95% CI: 0.43, 2.19), 0.55% (lag 8 weeks, 95% CI: 0.24, 0.85), 0.59% (lag 10 weeks, 95% CI: 0.16, 1.03), and 5.80% (lag 15 weeks, 95% CI: 2.96, 8.72), respectively. Our results indicated that ambient air pollutants were positively correlated with PTB risk, suggesting that decreasing outdoor air pollutant concentrations may help to reduce the burden of tuberculosis in countries with a high burden of tuberculosis and air pollution.

The Paradox of Predictability Provides a Bridge Between Micro- and Macroevolution.

The relationship between the evolutionary dynamics observed in contemporary populations (microevolution) and evolution on timescales of millions of years (macroevolution) has been a topic of considerable debate. Historically, this debate centers on inconsistencies between microevolutionary processes and macroevolutionary patterns. Here, we characterize a striking exception: emerging evidence indicates that standing variation in contemporary populations and macroevolutionary rates of phenotypic divergence are often positively correlated. This apparent consistency between micro- and macroevolution is paradoxical because it contradicts our previous understanding of phenotypic evolution and is so far unexplained. Here, we explore the prospects for bridging evolutionary timescales through an examination of this "paradox of predictability." We begin by explaining why the divergence-variance correlation is a paradox, followed by data analysis to show that the correlation is a general phenomenon across a broad range of temporal scales, from a few generations to tens of millions of years. Then we review complementary approaches from quantitative-genetics, comparative morphology, evo-devo, and paleontology to argue that they can help to address the paradox from the shared vantage point of recent work on evolvability. In conclusion, we recommend a methodological orientation that combines different kinds of short-term and long-term data using multiple analytical frameworks in an interdisciplinary research program. Such a program will increase our general understanding about how evolution works within and across timescales.

Learning the feature distribution similarities for online time series anomaly detection.

Identifying anomalies in multi-dimensional sequential data is crucial for ensuring optimal performance across various domains and in large-scale systems. Traditional contrastive methods utilize feature similarity between different features extracted from multidimensional raw inputs as an indicator of anomaly severity. However, the complex objective functions and meticulously designed modules of these methods often lead to efficiency issues and a lack of interpretability. Our study introduces a structural framework called SimDetector, which is a Local-Global Multi-Scale Similarity Contrast network. Specifically, the restructured and enhanced GRU module extracts more generalized local features, including long-term cyclical trends. The multi-scale sparse attention module efficiently extracts multi-scale global features with pattern information. Additionally, we modified the KL divergence to suit the characteristics of time series anomaly detection, proposing a symmetric absolute KL divergence that focuses more on overall distribution differences. The proposed method achieves results that surpass or approach the State-of-the-Art (SOTA) on multiple real-world datasets and synthetic datasets, while also significantly reducing Multiply-Accumulate Operations (MACs) and memory usage.

Reconstructing the age-structured case count of COVID-19 from sentinel surveillance data in Japan: a modeling study.

OBJECTIVES: To reconstruct age-structured case counts of COVID-19 using sentinel reporting, which replaced universal reporting of COVID-19 from May 2023 in Japan. METHODS: Using COVID-19 sentinel data stratified by discrete age groups in selected prefectures and referring to universal case count data up to May 8, 2023, we fitted a statistical model to handle weekly growth rates as a function of age and time so as to convert sentinel data to case counts after cessation of universal reporting. RESULTS: The age distribution of cases in sentinel reporting was significantly biased toward younger age groups compared to universal reporting. When comparing the epidemic size of the 9th wave (May 8 to September 18, 2023) to the 8th wave (October 3, 2022 to April 10, 2023), using the wave-on-wave ratio of total cumulative sentinel cases led to a significant underestimation of the wave-on-wave in Tokyo (0.975, vs 1.461 by universal reporting) and Okinawa (1.299, vs 1.472). The estimates of growth rates, scaling factors between universal and sentinel cases, and expected universal case count showed robustness to changes in the ending week of the data period. CONCLUSIONS: Our model quantified COVID-19 dynamics, comparably to universal reporting that ended in May 2023, enabling detailed and up-to-date health burden analysis using sentinel reports. The cumulative incidence was greater than that suggested from sentinel data in Tokyo, Nara, and Okinawa. Per-population burdens among children were particularly high in Osaka and Nara, indicating a strong bias in sentinel reporting toward pediatric cases.

Mining Alzheimer's disease clinical data: reducing effects of natural aging for predicting progression and identifying subtypes.

Introduction: Because Alzheimer's disease (AD) has significant heterogeneity in encephalatrophy and clinical manifestations, AD research faces two critical challenges: eliminating the impact of natural aging and extracting valuable clinical data for patients with AD. Methods: This study attempted to address these challenges by developing a novel machine-learning model called tensorized contrastive principal component analysis (T-cPCA). The objectives of this study were to predict AD progression and identify clinical subtypes while minimizing the influence of natural aging. Results: We leveraged a clinical variable space of 872 features, including almost all AD clinical examinations, which is the most comprehensive AD feature description in current research. T-cPCA yielded the highest accuracy in predicting AD progression by effectively minimizing the confounding effects of natural aging. Discussion: The representative features and pathogenic circuits of the four primary AD clinical subtypes were discovered. Confirmed by clinical doctors in Tangdu Hospital, the plaques (18F-AV45) distribution of typical patients in the four clinical subtypes are consistent with representative brain regions found in four AD subtypes, which further offers novel insights into the underlying mechanisms of AD pathogenesis.

Effective weight optimization strategy for precise deep learning forecasting models using EvoLearn approach.

Time series analysis and prediction have attained significant attention from the research community in the past few decades. However, the prediction accuracy of the models highly depends on the models' learning process. In order to optimize resource usage, a better learning methodology, in terms of accuracy and learning time, is needed. In this context, the current research work proposes EvoLearn, a novel method to improve and optimize the learning process of neural-based models. The presented technique integrates the genetic algorithm with back-propagation to train model weights during the learning process. The fundamental idea behind the proposed work is to select the best components from multiple models during the training process to obtain an adequate model. To demonstrate the applicability of EvoLearn, the method is tested on the state-of-the-art neural models (namely MLP, DNN, CNN, RNN, and GRU), and performances are compared. Furthermore, the presented study aims to forecast two types of time series, i.e. air pollution and energy consumption time series, using the developed framework. In addition, the considered neural models are tested on two datasets of each time series type. From the performance comparison and evaluation of EvoLearn using a one-tailed paired T-test against the conventional back-propagation-based learning approach, it was found that the proposed method significantly improves the prediction accuracy.

NMR Dynamic View of the Stabilization of the WW4 Domain by Neutral NaCl and Kosmotropic Na2SO4 and NaH2PO4.

The Hofmeister series categorizes ions based on their effects on protein stability, yet the microscopic mechanism remains a mystery. In this series, NaCl is neutral, Na2SO4 and Na2HPO4 are kosmotropic, while GdmCl and NaSCN are chaotropic. This study employs CD and NMR to investigate the effects of NaCl, Na2SO4, and Na2HPO4 on the conformation, stability, binding, and backbone dynamics (ps-ns and µs-ms time scales) of the WW4 domain with a high stability and accessible side chains at concentrations ≤ 200 mM. The results indicated that none of the three salts altered the conformation of WW4 or showed significant binding to the four aliphatic hydrophobic side chains. NaCl had no effect on its thermal stability, while Na2SO4 and Na2HPO4 enhanced the stability by ~5 °C. Interestingly, NaCl only weakly interacted with the Arg27 amide proton, whereas Na2SO4 bound to Arg27 and Phe31 amide protons with Kd of 32.7 and 41.6 mM, respectively. Na2HPO4, however, bound in a non-saturable manner to Trp9, His24, and Asn36 amide protons. While the three salts had negligible effects on ps-ns backbone dynamics, NaCl and Na2SO4 displayed no effect while Na2HPO4 significantly increased the µs-ms backbone dynamics. These findings, combined with our recent results with GdmCl and NaSCN, suggest a microscopic mechanism for the Hofmeister series. Additionally, the data revealed a lack of simple correlation between thermodynamic stability and backbone dynamics, most likely due to enthalpy-entropy compensation. Our study rationalizes the selection of chloride and phosphate as the primary anions in extracellular and intracellular spaces, as well as polyphosphate as a primitive chaperone in certain single-cell organisms.

Leveraging Deep Learning for Time-Series Extrinsic Regression in Predicting the Photometric Metallicity of Fundamental-Mode RR Lyrae Stars.

Astronomy is entering an unprecedented era of big-data science, driven by missions like the ESA's Gaia telescope, which aims to map the Milky Way in three dimensions. Gaia's vast dataset presents a monumental challenge for traditional analysis methods. The sheer scale of this data exceeds the capabilities of manual exploration, necessitating the utilization of advanced computational techniques. In response to this challenge, we developed a novel approach leveraging deep learning to estimate the metallicity of fundamental mode (ab-type) RR Lyrae stars from their light curves in the Gaia optical G-band. Our study explores applying deep-learning techniques, particularly advanced neural-network architectures, in predicting photometric metallicity from time-series data. Our deep-learning models demonstrated notable predictive performance, with a low mean absolute error (MAE) of 0.0565, the root mean square error (RMSE) of 0.0765, and a high R2 regression performance of 0.9401, measured by cross-validation. The weighted mean absolute error (wMAE) is 0.0563, while the weighted root mean square error (wRMSE) is 0.0763. These results showcase the effectiveness of our approach in accurately estimating metallicity values. Our work underscores the importance of deep learning in astronomical research, particularly with large datasets from missions like Gaia. By harnessing the power of deep-learning methods, we can provide precision in analyzing vast datasets, contributing to more precise and comprehensive insights into complex astronomical phenomena.

An End-to-End Deep Learning Framework for Fault Detection in Marine Machinery.

The Industrial Internet of Things has enabled the integration and analysis of vast volumes of data across various industries, with the maritime sector being no exception. Advances in cloud computing and deep learning (DL) are continuously reshaping the industry, particularly in optimizing maritime operations such as Predictive Maintenance (PdM). In this study, we propose a novel DL-based framework focusing on the fault detection task of PdM in marine operations, leveraging time-series data from sensors installed on shipboard machinery. The framework is designed as a scalable and cost-efficient software solution, encompassing all stages from data collection and pre-processing at the edge to the deployment and lifecycle management of DL models. The proposed DL architecture utilizes Graph Attention Networks (GATs) to extract spatio-temporal information from the time-series data and provides explainable predictions through a feature-wise scoring mechanism. Additionally, a custom evaluation metric with real-world applicability is employed, prioritizing both prediction accuracy and the timeliness of fault identification. To demonstrate the effectiveness of our framework, we conduct experiments on three types of open-source datasets relevant to PdM: electrical data, bearing datasets, and data from water circulation experiments.

An Integrated Data Acquisition Approach for the Structural Health Monitoring and Real-Time Earthquake Response Assessment of a Retrofitted Adobe Church in Peru.

The structural health monitoring (SHM) of buildings provides relevant data for the evaluation of the structural behavior over time, the efficiency of maintenance, strengthening, and post-earthquake conditions. This paper presents the design and implementation of a continuous SHM system based on dynamic properties, base accelerations, crack widths, out-of-plane rotations, and environmental data for the retrofitted church of Kuñotambo, a 17th century adobe structure, located in the Peruvian Andes. The system produces continuous hourly records. The organization, data collection, and processing of the SHM system follows different approaches and stages, concluding with the assessment of the structural and environmental conditions over time compared to predefined thresholds. The SHM system was implemented in May 2022 and is part of the Seismic Retrofitting Project of the Getty Conservation Institute. The initial results from the first twelve months of monitoring revealed seasonal fluctuations in crack widths, out-of-plane rotations, and natural frequencies, influenced by hygrothermal cycles, and an apparent positive trend, but more data are needed to justify the nature of these actions. This study emphasizes the necessity for extended data collection to establish robust correlations and refine monitoring strategies, aiming to enhance the longevity and safety of historic adobe structures under seismic risk.

Effect of Temperature and Precipitation on Acute Appendicitis Incidence in Seoul: A Time Series Regression Analysis.

PURPOSE: This study aimed to investigate the relationship between meteorological factors, specifically temperature and precipitation, and the incidence of appendicitis in Seoul, South Korea. METHODS: Using data from the National Health Insurance Service spanning 2010-2020, the study analyzed 165,077 appendicitis cases in Seoul. Time series regression modeling with distributed-lag non-linear models was employed. RESULTS: Regarding acute appendicitis and daily average temperature, the incidence rate ratio (IRR) showed an increasing trend from approximately - 10 °C to 10 °C. At temperatures above 10 °C, the increase was more gradual. The IRR approached a value close to 1 at temperatures below - 10 °C and above 30 °C. Both total and complicated appendicitis exhibited similar trends. Increased precipitation was negatively associated with the incidence of total acute appendicitis around the 50 mm/day range, but not with complicated appendicitis. CONCLUSIONS: The findings suggest that environmental factors, especially temperature, may play a role in the occurrence of appendicitis. This research underscores the potential health implications of global climate change and the need for further studies to understand the broader impacts of environmental changes on various diseases.

Time-series domain adaptation via sparse associative structure alignment: Learning invariance and variance.

Domain adaptation on time-series data, which is often encountered in the field of industry, like anomaly detection and sensor data forecasting, but received limited attention in academia, is an important but challenging task in real-world scenarios. Most of the existing methods for time-series data use the covariate shift assumption for non-time-series data to extract the domain-invariant representation, but this assumption is hard to meet in practice due to the complex dependence among variables and a small change of the time lags may lead to a huge change of future values. To address this challenge, we leverage the stableness of causal structures among different domains. To further avoid the strong assumptions in causal discovery like linear non-Gaussian assumption, we relax it to mine the stable sparse associative structures instead of discovering the causal structures directly. Besides the domain-invariant structures, we also find that some domain-specific information like the strengths of the structures is important for prediction. Based on the aforementioned intuition, we extend the sparse associative structure alignment model in the conference version to the Sparse Associative Structure Alignment model with domain-specific information enhancement (SASA2 in short), which aligns the invariant unweighted spare associative structures and considers the variant information for time-series unsupervised domain adaptation. Specifically, we first generate the segment set to exclude the obstacle of offsets. Second, we extract the unweighted sparse associative structures via sparse attention mechanisms. Third, we extract the domain-specific information via an autoregressive module. Finally, we employ a unidirectional alignment restriction to guide the transformation from the source to the target. Moreover, we further provide a generalization analysis to show the theoretical superiority of our method. Compared with existing methods, our method yields state-of-the-art performance, with a 5% relative improvement in three real-world datasets, covering different applications: air quality, in-hospital healthcare, and anomaly detection. Furthermore, visualization results of sparse associative structures illustrate what knowledge can be transferred, boosting the transparency and interpretability of our method.

Airborne particulate matter measurement and prediction with machine learning techniques.

Air quality is a fundamental component of a healthy environment for human beings. Monitoring networks for air pollution have been established in numerous industrial zones. The data collected by the pervasive monitoring devices can be utilized not only for determining the current environmental condition, but also for forecasting it in the near future. This paper considers the applications of different machine learning methods for the prediction of the two most widely used quantities. Particulate matter (PM) with a diameter of 2.5 and 10 µm, respectively. The data are collected via a proprietary monitoring station, designated as the Ecolumn. The Ecolumn monitors a number of key parameters, including temperature, pressure, humidity, PM 1.0, PM 2.5, and PM 10, in a timely manner. The data were employed in the development of multiple models based on selected machine learning methods. The decision tree, random forest, recurrent neural network, and long short-term memory models were employed. Experiments were conducted with varying hyperparameters and network architectures. Different time scales (10 min, 1 h, and 24 h) were examined. The most optimal results were observed for the Long Short-Term Memory algorithm when utilizing the shortest available time spans (shortest averaging times). The decision tree and random forest algorithms demonstrated unexpectedly high performance for long averaging times, exhibiting only a slight decline in accuracy compared to neural networks for shorter averaging times. Recommendations for the potential applicability of the tested methods were formulated.