Seasonal Component Research Articles

Predicting rainfall using machine learning, deep learning, and time series models across an altitudinal gradient in the North-Western Himalayas

AbstractPredicting rainfall is a challenging and critical task due to its significant impact on society. Timely and accurate predictions are essential for minimizing human and financial losses. The dependence of approximately 60% of agricultural land in India on monsoon rainfall implies the crucial nature of accurate rainfall prediction. Precise rainfall forecasts can facilitate early preparedness for disasters associated with heavy rains, enabling the public and government to take necessary precautions. In the North-Western Himalayas, where meteorological data are limited, the need for improved accuracy in traditional modeling methods for rainfall forecasting is pressing. To address this, our study proposes the application of advanced machine learning (ML) algorithms, including random forest (RF), support vector regression (SVR), artificial neural network (ANN), and k-nearest neighbour (KNN) along with various deep learning (DL) algorithms such as long short-term memory (LSTM), bi-directional LSTM, deep LSTM, gated recurrent unit (GRU), and simple recurrent neural network (RNN). These advanced techniques hold the potential to significantly improve the accuracy of rainfall prediction, offering hope for more reliable forecasts. Additionally, time series techniques, including autoregressive integrated moving average (ARIMA) and trigonometric, Box-Cox transform, arma errors, trend, and seasonal components (TBATS), are proposed for predicting rainfall across the altitudinal gradients of India’s North-Western Himalayas. This approach can potentially revolutionise how we approach rainfall forecasting, ushering in a new era of accuracy and reliability. The effectiveness and accuracy of the proposed algorithms were assessed using meteorological data obtained from six weather stations at different elevations spanning from 1980 to 2021. The results indicate that DL methods exhibit the highest accuracy in predicting rainfall, as measured by the root mean squared error (RMSE) and mean absolute error (MAE), followed by ML algorithms and time series techniques. Among the DL algorithms, the accuracy order was bi-directional LSTM, LSTM, RNN, deep LSTM, and GRU. For the ML algorithms, the accuracy order was ANN, KNN, SVR, and RF. These findings suggest that altitude significantly affects the accuracy of the models, highlighting the need for additional weather stations in this mountainous region to enhance the precision of rainfall prediction.

Seasonality and Predictability of Hydrometeorological and Water Chemistry Indicators in Three Coastal Forested Watersheds

Forests are recognized for sustaining good water chemistry within landscapes. This study focuses on the water chemistry parameters and their hydrological predictability and seasonality (as a component of predictability) in watersheds of varying scales, with and without human (forest management) activities on them, using Colwell indicators for data collected during 2011–2019. The research was conducted in three forested watersheds located at the US Forest Service Santee Experimental Forest in South Carolina USA. The analysis revealed statistically significant (α = 0.05) differences between seasons for stream flow, water table elevation (WTE), and all water chemistry indicators in the examined watersheds for the post-Hurricane Joaquin period (2015–2019), compared to the 2011–2014 period. WTE and flow were identified as having the greatest influence on nitrogen concentrations. During extreme precipitations events, such as hurricanes or tropical storms, increases in WTE and flow led to a decrease in the concentrations of total dissolved nitrogen (TDN), NH4-N, and NO3-N+NO2-N, likely due to dilution. Colwell indicators demonstrated higher predictability (P) for most hydrologic and water chemistry indicators in the 2011–2014 period compared to 2015–2019, indicating an increase in the seasonality component compared to constancy (C), with a larger decrease in C/P for 2015–2019 compared to 2011–2014. The analysis further highlighted the influence of extreme hydrometeorological events on the changing predictability of hydrology and water chemistry indicators in forested streams. The results demonstrate the influence of hurricanes on hydrological behavior in forested watersheds and, thus, the seasonality and predictability of water chemistry variables within and emanating out of the watershed, potentially influencing the downstream ecosystem. The findings of this study can inform forest watershed management in response to natural or anthropogenic disturbances.

National Influenza Annual Report 2023-2024: A focus on influenza B and public health implications.

The 2023-2024 influenza epidemic saw the return of typical late-season influenza B circulation. The epidemic was declared in week 45 (week ending November 11, 2023) due to the predominant circulation of influenza A(H1N1) and peaked in week 52 (week ending December 30, 2023); however, as influenza A circulation decreased, influenza B detections and the percentage of tests positive increased, reaching its peak in week 14 (week ending April 6, 2024). Influenza B/Victoria dominated this wave of activity, contributing to the ongoing discussion about the apparent disappearance of influenza B/Yamagata. With the recommendation for the removal of influenza B/Yamagata lineages from the recommended seasonal influenza vaccine components, the influenza surveillance community is preparing for the possibility of a new seasonal pattern dominated by influenza B/Victoria circulation. This season, as a result of influenza B/Victoria's overwhelming predominance, younger age groups were primarily affected by the wave of influenza B activity. Over the course of the season, among all influenza B detections, 52% occurred in children aged 0-19 years. Among all influenza B-associated hospitalizations, 46.4% were in children aged 0-19 years, and the highest cumulative hospitalization rates for influenza B were among children younger than five years (n=37 per 100,000 population) and children between the ages of 5-19 years (n=15 per 100,000 population). Continued vigilance and surveillance around influenza B trends and epidemiology is required to contribute to effective epidemic preparedness.

An integrated approach for decomposing time series data into trend, cycle and seasonal components

ABSTRACT This study presents a methodology for decomposing time series data into trend, seasonal, and cyclical components using the moving linear model approach by Kyo and Kitagawa (Journal of Business Cycle Research, 19(3): 373-397, 2023). Our approach integrates seasonal adjustment and decomposition into a single framework. We evaluated our approach with two case studies: examining daily COVID-19 case data and the Index of Industrial Production (IIP) in Japan, comparing the results to seasonally adjusted IIP data. Performance metrics included the discrimination power index and the variance of the adjusted cyclical component. Our findings show that our method effectively extracts business cycle information, achieving higher discrimination power and greater adjusted variance compared to seasonally adjusted IIP data, highlighting the superior performance of our integrated seasonal adjustment method. We compared the proposed approach with a state-space modelling method by introducing an overall stability as a new indicator. The results demonstrated the stability of the estimations obtained with our proposed method.

On several properties of a novel class of generalized Humbert autoregressive moving average process

The main objective of this paper is to introduce a new class of time series that extends a specific fractional models, including fractional autoregressive integrated moving average, Gegenbauer autoregressive moving average and Humbert autoregressive moving average (HARMA). The generalized HARMA is constructed using an orthogonal polynomial called generalized Humbert polynomial. The process is constructed to enhance the ability to capture complex patterns such as long memory and seasonal components. In this study, we present several properties including stationarity, invertibility, spectral density, and autocovariance function. In the simulation study, we check the validity of the theoretical findings.

Analysis of Google users' interests in information about pacifiers: an ecological digital study.

To determine the interests of Google users from different countries about pacifier-related information over time, through a quantitative and qualitative analysis. This infodemiology longitudinal retrospective ecological study investigated pacifier-related computational metadata from 22 countries using Google Trends platform. Initially, Relative Search Volume (RSV) time series data and their respective main queries were collected employing the topic "pacifier - subject" as the search strategy from January 2004 to June 2024, considering all categories. Data were then analyzed through seasonal decomposition to determine the trend, seasonal, and residual components of the collected time series using Python 3 programming libraries in a Google Collaboratory interface. Additionally, Mann-Kendall tests were performed to access the significance level of trends observed from the seasonal decomposition. Finally, the queries were qualitatively evaluated to identify the main subjects related to users' interests. Statistical significance was determined with p values < 0.05. In general, RSV time series curves exhibited upward trends in all analyzed countries over the years, with significance levels demonstrated by the Mann-Kendall tests. Moreover, queries often centered around types of pacifiers, encompassing searches specifically for certain commercial brands. Some searches concerning newborn and baby care were also observed, particularly highlighting breastfeeding practices. Google users have shown an increasing interest in pacifier-related topics. These data can help understand the needs and concerns of parents and caregivers regarding pacifier use, ultimately supporting the development of effective educational strategies.

Exceptionally Strong Equatorial Intermediate Current Events in the Indian Ocean Associated with Climate Modes

Abstract Knowledge of the effects of climate modes on the equatorial intermediate current (EIC) remains limited. This paper investigates exceptional events of the EIC in the Indian Ocean and their relationships with climate modes at various time scales by using observations, reanalysis outputs, and a continuously stratified linear ocean model (LOM). A mooring at 80°E from 2015 to 2019 revealed four exceptionally strong EIC events, occurring in 2015 July–August (JA), 2016 January–February (JF), 2016 JA, and 2019 JF. Component analysis revealed that these exceptional events are attributed to the co-occurrence of the seasonal components peaking during JF and JA, as well as the larger current anomalies associated with intraseasonal and interannual components. In the intraseasonal band, the Madden–Julian oscillation (MJO) generates a significant EIC anomaly through a 40–50-day process involving equatorial waves. The MJO exerts a substantial effect when the amplitude of the MJO index exceeds 1 and the oscillation is in phase 4. In the interannual band, El Niño–Southern Oscillation and the Indian Ocean dipole (IOD) can each independently contribute to the EIC anomaly. This contribution initiates during the mature phase and persists throughout the subsequent year. Concurrent occurrences of these effects can lead to larger interannual anomalies. Notably, El Niño–positive IOD (El Niño–pIOD) and La Niña–negative IOD (La Niña–nIOD) synergies are most pronounced in August and in January of the following year. Significance Statement Four exceptional equatorial intermediate current (EIC) events characterized by strong eastward velocities are observed in the Indian Ocean during 2015–19. This study explored their underlying dynamics and their relationships with climate modes. Climatologically, the EIC has seasonal cycles with peaks in January–February and July–August, which is helpful in producing exceptional EIC events. However, the occurrence of these events is attributed to the intensity of the intraseasonal and interannual variability that occurs around the seasonal cycle peaks. This study revealed that climate modes, including the Madden–Julian oscillation (MJO), El Niño–Southern Oscillation (ENSO), and Indian Ocean dipole (IOD), contribute to exceptional EIC events in the intraseasonal and interannual bands. Notably, their influence is comparable, albeit contingent upon specific conditions.

High-frequency realized stochastic volatility model

A new high-frequency realized stochastic volatility model is proposed. Apart from the standard daily-frequency stochastic volatility model, the high-frequency stochastic volatility model is fit to intraday returns by extensively incorporating intraday volatility patterns. The daily realized volatility calculated using intraday returns is incorporated into the high-frequency stochastic volatility model by considering the bias in the daily realized volatility caused by microstructure noise. The volatility of intraday returns is assumed to consist of the autoregressive process, the seasonal component of the intraday volatility pattern, and the announcement component responding to macroeconomic announcements. A Bayesian method via Markov chain Monte Carlo is developed for the analysis of the proposed model. The empirical analysis using the 5-minute returns of E-mini S&P 500 futures provides evidence that our high-frequency realized stochastic volatility model improves in-sample model fit and volatility forecasting over the high-frequency stochastic volatility model.

Tobacco Free Advent Calendar: Evaluating innovative strategies in Public Health digital campaigns

Abstract Introduction Two editions, 2022 and 2023, of the Tobacco Free Advent Calendar (TFAC) built a large network of Organizations, Universities, Institutions, individual experts and medical students across Europe, coordinated by ENSPNext (European Network for Smoking and Tobacco Prevention Youth Group), to advocate for Youth and for Public Health. This study has 2 objectives: 1. Monitor the performance of TFAC 2023 vs TFAC 2022 for 31 days 2. Evaluate improvements in reach and engagement made for TFAC 2023. Methods For 24 days, TFAC 2023 featured thematic messages, embedded fact sheets, quizzes, videos and call-to-action buttons. New content was accessible daily starting on the first day of advent, ending on Christmas. TFAC was available in English and Armenian. AI was used to assist for various tasks and challenges during development and research. Comparative and Time Series Analysis were conducted in Excel and IBM SPSS26. Results TFAC 2023 was accessed 3764 times (8.8% increase), the doors were opened 1506 times (3.9% decrease), Call to action buttons were clicked 261 times (169% increase). For english version a strong variation in reach was observed, but not in engagement which remained relatively constant. In Armenian version both reach and engagement varied in 2023 vs 2022. A seasonal component was identified for both versions, particularly an increase at weekends and a decrease on weekdays. Similar to 2022, the reach and engagement were higher at the start and lower as time passed, although in 2023 the reach and engagement remained higher and more constant over time. Completion rates of interactive materials ranged between 21.4% to 100% in the English version and from 7.1% to 100% in the Armenian version. No interactive material remained unanswered. Conclusions Decrease of translations didn’t lower the reach and engagement of TFAC. Engagement with TFAC 2023 was higher, possibly due to the implication of students organizations and collaboration with influencers. Key messages • Targeted dissemination increased the reach and engagement of TFAC 2023 versus TFAC 2022. • Members of the target group should be recruited for being members in the projects affecting them to increase engagement and impact of the campaign.

A comparative investigation of monsoon active and break events over the western North Pacific

Abstract This study identifies 86 active events and 66 break events of the western North Pacific summer monsoon (WNPSM) from 1979 to 2020. These active and break events exhibit sharp contrast in large-scale convection and circulation fields. During the active period, deep convection almost covers the entire WNP domain (0°–25°N, 120°–170°E), accompanied by the strengthening of a local monsoon trough and monsoon westerlies extending from the Arabian Sea to the WNP. In contrast, during the break period, along with weakened monsoon westerlies, the subtropical high replaces the monsoon trough to dominate the WNP domain, leading to disappearance of deep convection. Results about the convection evolution at multiple timescales indicate that active/break events are primarily contributed to by the wet/dry phases of intraseasonal oscillations (ISOs). The phase transitions of ISOs are nearly symmetric between active and break events. However, these two types of events present notable asymmetric features in convection anomalies relative to climatology, which can be attributable to the modulation of the seasonal mean component. Specifically, both active and break events correspond to anomalously strong seasonal mean convection, which amplifies (weakens) the convection enhancement (suppression) during the active (break) period and results in normal (enhanced) convection during adjacent periods. The preferred occurrence of active and break events in strong summer monsoons is because related mean backgrounds, including increased low-level positive vorticity, specific humidity, and easterly vertical shear, facilitate the intensification of ISOs. These mean backgrounds are likely induced by SST cooling in the north Indian Ocean and warming in the equatorial central Pacific during the simultaneous summer.

Forecasting to support EMS tactical planning: what is important and what is not.

Forecasting emergency medical service (EMS) call volumes is critical for resource allocation and planning. The development of many commercial and free software packages has made a variety of forecasting methods accessible. Practitioners, however, are left with little guidance on selecting the most appropriate method for their needs. Using 5 years of data from 3 cities in Alberta, we compute exponential smoothing and benchmark forecasts for 8-hour periods for each ambulance station catchment area and with a forecast horizon of two weeks-a spatio-temporal resolution appropriate for tactical planning. The methods that we consider differ on three spectra: the number and type of time-series components, whether forecasts are computed individually or jointly, and the way in which forecasts at a specific resolution are converted to forecasts at the resolution of interest. We find that it is important to include a weekly seasonal component when forecasting EMS demand. Multiplicative seasonality, however, shows no benefit over additive seasonality. Adding other time-series components (e.g., trend, ARMA errors, Box-Cox transformation) does not improve performance. Spatial resolutions of station catchment area and lower, and temporal resolution of 4-24 hours perform similarly. We adapt an existing hierarchical forecasting framework to a two-dimensional spatio-temporal hierarchy, but find that hierarchical reconciliation of forecasts does not improve performance at the forecast resolution of interest for tactical planning. Neither does jointly forecasting time series. We show that added complexity does not materially improve forecasting performance. The simple methods that we find perform well are easy to implement and interpret, making implementation in practice more likely. In a simulation study we alter the empirical weekly patterns and demonstrate how extreme differences between the weekly seasonality patterns of different regions cause hierarchically-reconciled bottom-up approaches to outperform top-down approaches.

Spatiotemporal variations and driving factors of evapotranspiration in the Yunnan-Guizhou Plateau from 2003 to 2020

ABSTRACT Evapotranspiration (ET) is vital for the Earth's energy and water balance, particularly influenced by global climate change. The Yunnan-Guizhou Plateau (YGP), characterized by abundant water resources and intricate terrain, has been a subject of study. However, previous research often overlooked intra-annual climate variations in ET. This study employed high-spatiotemporal resolution ET data from 2003 to 2020 to quantitatively analyze the spatiotemporal characteristics of ET on the YGP. The annual ET showed an increasing trend of 0.18 mm/year, with monthly ET increases in January, March, November, and December, mainly influenced by vegetation transpiration, which accounts for 56% of ET. Breakpoints in ET trends and seasonal components occurred in January 2007 and June 2018. The geodetector model assessed the impact of 15 driving factors on ET, with net radiation and vegetation index playing dominant roles with q-values of 0.29 and 0.24. Factor impacts varied seasonally, with greater influence in the dry season (q-value of 0.53 for net radiation in January) and lesser in the rainy season (q-value of 0.08 in August). Pearson correlation analysis indicated that different driving factors influenced ET in different months. These findings enhance understanding of plateau ET responses to climate change mechanisms.

The evaluation of competition effect on rail fares using the difference-in-difference method through symmetric and lagged spans

Long-distance railway markets in Europe have become more competitive since the EU liberalisation packages. Countries like Austria, Czechia, Sweden, and Italy have experienced sustained competition, providing valuable insights into how open-access competition works. Italy's non-subsidised high-speed competition is especially interesting to study. While some researchers have studied the minimum price changes before and after a newcomer enters a route, more is needed about the duration of the effect and the impact on all fare types. Our paper estimates the competition's impact by observing the price dynamics of the incumbent company (Trenitalia) for different fare types over an extended period. We examine the Turin-Milan-Venice route where the competitor (Italo) entered the market during observation. We distinguish between different fare types through detailed descriptive analyses of the incumbent company's price behaviour. Using the difference-in-difference method, we obtain estimates of the effect for different time spans and check their robustness.Two methods are used to design difference-in-difference models: a “classical approach” that compares symmetrical intervals before and after entry and another approach based on the comparison of lagged intervals. The second method compares the specific span with the corresponding one, a year before to exclude the seasonal component. Our findings reveal significant price-reduction effects compared to the price dynamics in other market routes where competition had been introduced.

Temporal Trends in the Prevalence of Child Undernutrition in China From 2000 to 2019, With Projections of Prevalence in 2030: Cross-Sectional Analysis.

Although the problem of malnutrition among children in China has greatly improved in recent years, there is a gap compared to developed countries, and there are differences between provinces. Research on long-term comprehensive trends in child growth failure (CGF) in China is needed for further improvement. The purpose of this study was to examine trends in stunting, wasting, and underweight among children younger than 5 years in China from 2000 to 2019, and predict CGF till 2030. We conducted a cross-sectional analysis using data from the local burden of disease (LBD) database. Using Joinpoint Regression Software, we examined trends in CGF among children younger than 5 years in China from 2000 to 2019, and predicted the trends of prevalence in 2030, using the Holt-Winters model with trends but without seasonal components. The assessment was performed with Stata 17 (StataCorp). Data were analyzed from October 17, 2023, to November 22, 2023. In 2019, the prevalences of stunting, wasting, and underweight decreased to 12%, 3%, and 4%, respectively (decreases of 36.9%, 25.0%, and 42.9%, respectively, compared with the values in 2000). The prevalence of CGF decreased rapidly from 2000 to 2010, and the downward trend slowed down after 2010. Most provinces had stagnated processes of trends after 2017. The age group with the highest stunting prevalence was children aged 1 to 4 years, and the highest prevalence of wasting and underweight was noted in early neonatal infants. From 2000 to 2019, the prevalence of CGF declined in all age groups of children. The largest relative decrease in stunting and underweight was noted in children aged 1 to 4 years, and the largest decrease in wasting was noted in early neonatal infants. The prevalences of stunting, wasting, and underweight in China are estimated to decrease to 11.4%, 3.2%, and 4.1%, respectively, by 2030. China has nationally met the World Health Organization's Global Nutrition Targets for 2030 for stunting but not for wasting. This study provides data on the prevalence and trends of CGF among children younger than 5 years and reports declines in CGF. There remain areas with slow progress in China. Most units have achieved the goal for stunting prevalence but not wasting prevalence.

Spatiotemporal Characterization and Driving Factors of Fine Particulate Matter and Its Chemical Components in the Huaihe River Basin

According to the data sets of fine particulate matter （PM2.5） and its components in 35 cities in the Huaihe River Basin from 2015 to 2021, the temporal and spatial distribution patterns of pollutants were analyzed. The influence of meteorological factors on PM2.5 concentrations was examined using a random forest model. The original series of PM2.5, sulfate （SO42-）, nitrate （NO3-）, ammonium salt （NH4+）, organic matter （OM）, and black carbon （BC） were rebuilt using KZ （Kolmogorov-Zurbenko） filtering and multiple linear regression （MLR） to quantify the effects of meteorological conditions. The results demonstrated that from 2015 to 2021, the declining rates of PM2.5, SO42-, NO3-, NH4+, OM, and BC in the Huaihe River Basin were 4.71, 0.99, 1.05, 0.77, 1.01, and 0.19 μg·（m3·a）-1, respectively. The high mass concentrations of PM2.5 and its components were concentrated in the central and western regions of the HRB, whereas those in coastal and southern cities were lower. The variance contributions of the short-term, seasonal, and long-term components of PM2.5 to the original PM2.5 sequences in 35 cities were 51.6%, 35.9%, and 7.0%, respectively. The PM2.5 in coastal cities were more affected by the short-term components. The meteorological conditions were unfavorable for PM2.5 reduction in the HRB from 2015 to 2018, whereas the meteorological conditions supported the PM2.5 decrease from 2019 to 2021. From 2015 to 2021, the contribution rates of meteorological conditions to the long-term component reductions of PM2.5, SO42-, NO3-, NH4+, OM, and BC were 28.3%, 29.1%, 31.0%, 29.3%, 27.8%, and 28.6%, respectively. The contribution rates of meteorological conditions to the long-term PM2.5 reduction were 43.4%, 25.6%, 25.5%, and 20.6% in the HRB cities in Anhui, Shandong, Jiangsu, and Henan Provinces, respectively. With the decrease in PM2.5 concentration in the HRB, the sulfur oxidation rate （SOR） increased significantly, while the nitrogen oxide oxidation rate （NOR） changed little.

Disentangling seasonal and annual precipitation signals in the tropics over the Holocene: Insights from δD, alkanes and GDGTs

Rainfall seasonality in the tropics has a substantial impact on both ecosystems and human livelihoods. Yet, reconstructions of past rainfall variability have so far generally been unable to differentiate between annual and seasonal precipitation changes. Past variations in seasonality are therefore largely unknown. Here, we disentangle hydrogen isotopic (δD) signals from terrestrial leaf waxes and algae in an 8000-year peat core from Sumatra, which reflect annual versus wet season rainfall signals, respectively. We validate these results using lipid biomarkers by reconstructing vegetation dynamics via n-alkane distributions and peatland hydrological conditions using glycerol dialkyl glycerol tetraethers (GDGTs), as well as biomass burning using levoglucosan concentrations in the core. Finally, we compare our proxy results to a transient climate model simulation (MPI-ESM1.2) to identify the mechanism for seasonality changes. We find that algal δD indicates stronger Indonesian-Australian Summer Monsoon (IASM) precipitation in the Mid-Holocene, between 8 and 4.2 cal ka BP. A period of alternating flooding, droughts and wildfires is reconstructed between 6 and 4.2 cal ka BP, implicating very strong monsoonal precipitation and drying out and burning during a longer and intensified dry season. We attribute this strong rainfall seasonality in the Mid-Holocene mainly to orbitally forced insolation seasonality and a strengthened IASM, consistent with the modeling results. In terms of annual rainfall, terrestrial plant δD, vegetation composition and GDGTs all indicate wetter conditions peaking between 3 and 4.5 cal ka BP, preceded by drier conditions, followed by drastic and rapid drying in the late Holocene from around 2.8 cal ka BP. Our multiproxy annual precipitation reconstruction thereby indicates the wettest overall conditions approximately 1500–2000 years later than a nearby speleothem δ18O record, which instead follows the seasonally biased algal δD in our record. We, therefore, hypothesize that speleothem reconstructions over the Holocene in parts of the tropics with low but significant seasonality may carry a stronger seasonal component than previously suggested. The data presented here contribute with new insights on how isotopic rainfall proxies in the tropics can be interpreted. Our findings resolve the seasonal versus annual components of Holocene rainfall variability in the Indo-Pacific Warm Pool region, highlighting the importance of considering seasonality in rainfall reconstructions.

Spatial-temporal analysis and trend prediction of regional crop disease based on electronic medical records

Intelligent diagnosis of individual crop diseases has matured. How to understand the evolution patterns and predict regional disease trends remains a significant challenge. Plant Electronic Medical Records (PEMRs) offer valuable spatial-temporal characteristics about crop diseases, presenting a new opportunity for predicting the occurrence of regional diseases. In this study, we used a large prescription database from Beijing (2018-2021) to reframe regional disease prediction as a time series forecasting task. Firstly, to analyze spatial-temporal evolution patterns, we use ArcGIS to extract key information and identify potential connections between different disease occurrence points. Then, we developed a novel deep learning combined model SV-CBA, which combines Seasonal and Trend decomposition (STL) with Variational Mode Decomposition (VMD) to identify trend, seasonal, and residual components, and re-decomposes the residuals. STL-VMD can capture long-term trends and periodic variations while managing nonlinear and volatile characteristics. The CNN-BiLSTM-Attention model calculates disease trends by linearly integrating predictions of each sub-series. To reduce computational complexity while maintaining predictive performance, we propose an improved simplified attention mechanism. Our model demonstrates superior performance in both comparative and ablation experiments using the PEMRs dataset, outperforming numerous other models. This study provides accurate disease trend predictions, aiding farmers and regional managers in agricultural production management.

A Novel Approach for Temperature Forecasting in Climate Change Using Ensemble Decomposition of Time Series

AbstractThis paper presents FORSEER (Forecasting by Selective Ensemble Estimation and Reconstruction), a novel methodology designed to address temperature forecasting under the challenges inherent to climate change. FORSEER integrates decomposition, forecasting, and ensemble methods within a modular framework. This methodology decomposes the time series into trend, seasonal, and residual components. Subsequently, multiple optimized forecast models are applied to each component. These component models are then carefully weighted and combined through an ensemble process to generate a final robust forecast. Experimental results demonstrate that FORSEER is an efficient computational forecasting methodology for complex climate time series. Furthermore, we show that FORSEER has an equivalent forecasting performance to the M4 competition champion SMYL method for temperature series. Besides, the proposed methodology has less computational complexity than SMYL, making it a more accessible and scalable option. FORSEER's modular architecture also allows flexibility when substituting techniques depending on the context of the problem, facilitating the parallel execution of independent tasks and resulting in a strategy adaptable to multiple contexts.

Zoonotic outbreak risk prediction with long short-term memory models: a case study with schistosomiasis, echinococcosis, and leptospirosis

BackgroundZoonotic infections, characterized with huge pathogen diversity, wide affecting area and great society harm, have become a major global public health problem. Early and accurate prediction of their outbreaks is crucial for disease control. The aim of this study was to develop zoonotic diseases risk predictive models based on time-series incidence data and three zoonotic diseases in mainland China were employed as cases.MethodsThe incidence data for schistosomiasis, echinococcosis, and leptospirosis were downloaded from the Scientific Data Centre of the National Ministry of Health of China, and were processed by interpolation, dynamic curve reconstruction and time series decomposition. Data were decomposed into three distinct components: the trend component, the seasonal component, and the residual component. The trend component was used as input to construct the Long Short-Term Memory (LSTM) prediction model, while the seasonal component was used in the comparison of the periods and amplitudes. Finaly, the accuracy of the hybrid LSTM prediction model was comprehensive evaluated.ResultsThis study employed trend series of incidence numbers and incidence rates of three zoonotic diseases for modeling. The prediction results of the model showed that the predicted incidence number and incidence rate were very close to the real incidence data. Model evaluation revealed that the prediction error of the hybrid LSTM model was smaller than that of the single LSTM. Thus, these results demonstrate that using trending sequences as input sequences for the model leads to better-fitting predictive models.ConclusionsOur study successfully developed LSTM hybrid models for disease outbreak risk prediction using three zoonotic diseases as case studies. We demonstrate that the LSTM, when combined with time series decomposition, delivers more accurate results compared to conventional LSTM models using the raw data series. Disease outbreak trends can be predicted more accurately using hybrid models.

Displacement Interval Prediction Method for Arch Dam with Cracks: Integrated STL, MF-DFA and Bootstrap

Displacement prediction models based on measured data have been widely applied in structural health monitoring. However, most models neglect the particularity of displacement monitoring for arch dams with cracks, nor do they thoroughly analyze the non-stationarity and uncertainty of displacement. To address this issue, the influencing factors of displacement were first considered, with crack opening displacement being incorporated into them, leading to the construction of the HSCT model that accounts for the effects of cracks. Feature selection was performed on the factors of the HSCT model utilizing the max-relevance and min-redundancy (mRMR) algorithm, resulting in the screened subset of displacement influence factors. Next, displacement was decomposed into trend, seasonal, and remainder components applying the seasonal-trend decomposition using loess (STL) algorithm. The multifractal characteristics of these displacement components were then analyzed by multifractal detrended fluctuation analysis (MF-DFA). Subsequently, displacement components were predicted employing the convolutional neural network-long short-term memory (CNN-LSTM) model. Finally, the impact of uncertainty factors was quantified using prediction intervals based on the bootstrap method. The results indicate that the proposed methods and models are effective, yielding satisfactory prediction accuracy and providing scientific basis and technical support for the health diagnosis of hydraulic structures.