Prediction Intervals Research Articles

Past, current, and future trends in the prevalence of primary sclerosing cholangitis and inflammatory bowel disease across England (2015–2027): a nationwide, population-based study

Primary sclerosing cholangitis (PSC) is one of the leading indications for liver transplantation in Europe, and a major risk factor for cancer in inflammatory bowel disease (IBD). However, it is not known how the epidemiology of PSC will change as that of IBD evolves. The aim of this study is to provide nationwide statistics on the past and current prevalence of PSC and IBD across England, and forecast how this is likely to change over time. We accessed and analysed a nationwide population-based administrative healthcare registry, which houses prospectively accrued data since April 1st 2001. In so doing, the past and current prevalence of PSC-IBD and IBD alone was determined among 18-60-year-olds in England, alongside average annual percentage change rates (AAPC), between the 1st of January 2015 and 2020. Past and current prevalence data, alongside trends in incidence and event-free survival rates, were then used to forecast future prevalence between 2021 and2027. In 2015, the prevalence of PSC with prior IBD diagnosis was 5.0 per 100,000 population, rising to 5.7 when including those with IBD diagnosed after PSC. In 2020, prevalence increased to 7.6 (8.6 accounting for IBD developing after PSC), yielding an AAPC of 8.8. In 2027, PSC-IBD prevalence is forecast to be 11.7 (95% prediction interval [PI]: 10.8-12.7), and 13.3 when accounting for IBD developing after PSC (AAPC: 6.4; 95% PI: 5.3-7.5). Comparatively, the prevalence of IBD alone rose among 18-60-year-olds from 384.3 in 2015 to 538.7 in 2020 (AAPC 7.0), and forecast to increase to 742.5 by 2027 (95% PI: 736.4-748.0; AAPC: 4.7, 95% PI: 4.6-4.8). The rate of growth in PSC-IBD is predicted to exceed IBD-alone. Further research is needed to understand changes in disease epidemiology, including aetiological drivers of developing (invariably progressive) liver disease in IBD, and the implications of rising case burden on health care resources. This study was supported by an unrestricted grant provided by Gilead Sciences.

Parental Inflammatory Bowel Disease with Child Autism: A Comprehensive Review and Meta-Analysis.

Inflammatory bowel disease (IBD) has been shown to be connected to a greater possibility of neurologically developed problems, such as autism spectrum disorders (ASDs). However, the proof linking parental IBD with ASD in offspring is inconclusive. Thus, we carried out a meta-analysis and comprehensive review to elucidate such linking. Prior research was identified through reviewing multiple internet-based sources, including Cochrane, Web of Knowledge, Embase, CINAHL, PubMed, and PsycINFO, from 1960 to December 2022. Pooled relative risks (RRs) and corresponding 95% confidence intervals (CIs) were determined employing random-effects models, in spite of the I2 statistic measurement of heterogeneity. Prediction intervals (PIs) have been presented to allow for more useful inferences and to indicate the range of genuine effects that might be predicted in future scenarios. Six studies (two case-control studies and four cohort studies) involving 3,200,199 participants were incorporated into the meta-analysis. The pooled RRs of ASDs among offspring of IBD parents were 1.15 (95% CI, 0.92 to 1.45, P = 0.226; I2 = 81.4%, P = 0.003; PI, 0.53-2.62), indicating no significant connection between parental IBD and the likelihood of ASDs in children. Type of IBD, and sex both also yielded no statistically significant results according to subgroup analysis. Our meta-analysis does not provide evidence that parental IBD is connected with the elevated likelihood of ASDs in their children. To confirm these results and understand their underlying mechanisms, additional research with larger sample sizes and improved study designs is required.

Which are the clinical consequences of Molar Incisor hypomineralization (MIH) in children and adolescents? Systematic review and meta-analysis.

This systematic review was conducted to address the following research question: "What are the clinical consequences (outcome) of Molar Incisor Hypomineralization (MIH) (exposure) in children and adolescents (population/patient)?". After defining the strategy, a search was performed in different databases (MEDLINE via Pubmed, Cochrane Library, BBO, LILACS, Scopus, Web of Science, Embase) and Grey literature in August 2023. Cross-sectional observational studies that identified clinical consequences of MIH (dental caries, post-eruptive structural loss, atypical restorations, hypersensitivity and tooth extraction) were included. The risk of bias was assessed following the Joanna Briggs Institute protocol for cross-sectional studies. Meta-analyses were conducted for each outcome, taking into account the number of patients and teeth. The effect measure considered was the prevalence; random-effects model was adopted. Heterogeneity was assessed using I2 statistics and prediction intervals (PI). A total of 903 studies were identified; 41 were selected for qualitative analysis and 38 for quantitative analysis. Twenty eight studies were classified as presenting uncertain risk of bias, 11 as low risk and 3 as high risk of bias. The prevalence levels, ranked from highest to lowest and considering the tooth and patient units, respectively, were: caries lesions (0.252 - 95% CI 0.158-0.375; 0.512 - 95% CI 0.385-0.639); hypersensitivity (0.286 - 95% CI 0.190-0.407; 0.417 - 95% CI 0.197-0.674), post-eruptive fracture (0.125 - 95% CI 0.099-0.158; 0.257 - 95% CI 0.145-0.412); atypical restorations (0.048 - 95% CI 0.030-0.077; 0.167 - 95% CI 0.096 - 0.274); tooth extraction (0.012 - 95% CI 0.007-0.019; 0.090 - 95% CI 0.019 - 0.331). All meta-analyses resulted in heterogeneity greater than 85%, with the exception of the outcome "tooth extraction" according to the tooth unit (I2 = 57.83). This heterogeneity may be attributed to factors such as differences in the location where the study was realized, the socioeconomic conditions of the studied population, the asymmetric nature of MIH, and patient age. The most common consequences of MIH are caries lesions, hypersensitivity, and post-eruptive breakdown. (PROSPERO:CRD42020201410).

Storage of weights and retrieval method (SWARM) approach for neural networks hybridized with conformal prediction to construct the prediction intervals for energy system applications

Storage of weights and retrieval method (SWARM) approach for neural networks hybridized with conformal prediction to construct the prediction intervals for energy system applications

Projecting exceedance of juvenile salmonid thermal maxima in streams under climate change: A crosswalk from lab experiments to riparian restoration

Abstract Concern over rising water temperatures for freshwater ectotherms has led to application of experimentally derived thermal thresholds to stream temperature models for estimation of streams at high risk of exceeding thermal thresholds under current and future climate conditions. We optimised an approach that links field‐relevant thermal maxima experiments to corresponding stream temperature models and identifies opportunities to reduce stream temperatures through riparian tree growth. We conducted a thermal maxima experiment on cold‐water adapted juvenile Chinook (Oncorhynchus tshawytscha) and coho salmon (Oncorhynchus kisutch) that reflected natural temperatures by using incremental temperature ramping with diel fluctuations (IT‐Dmax) and refit a regionally specific stream temperature model for British Columbia, Canada to directly relate to lab‐derived thresholds. Salmon‐bearing streams across British Columbia were categorised by threshold exceedance risk (i.e., low, moderate, high, severe) based on risk tolerance scenarios for management decision making (i.e., considering a range of prediction intervals [PIs] and stream thermal sensitivities to air temperature). We linked these results directly to riparian management and restoration actions by estimating the potential for riparian tree growth to reduce threshold exceedance risk. Lab‐derived IT‐Dmax endpoints were consistently 24°C (based on the median value between the 7‐day average of the daily maximum and the mean weekly average temperature) across acclimation treatments for both species. Under current conditions, most stream reaches (99.6% using an intermediate risk tolerance scenario) were below the thermal threshold; streams with moderate to high risk of threshold exceedance were projected to increase from 0.4% to 1.5% (total linear stream length of 6,929 km) by end of century. The risk‐averse (high sensitivity, 75% PI) and risk‐tolerant (low sensitivity, 95% PI) scenarios differed by 1,107 km of streams predicted to have a moderate to severe risk of exceeding thresholds by end of century. Maximal riparian tree growth was predicted to shift 670 km of streams from moderate to low exceedance risk under end‐of‐century conditions and intermediate risk tolerance, showing the potential for mitigation from thermal impacts. Our integrative approach tackled several key considerations in identifying streams at high temperature risk for freshwater ectotherms that to date have not been addressed. Furthermore, we showcased the delineation of spatially comprehensive estimates that identify where management mitigation and a specific restoration activity may be most needed to reduce impacts of rising temperatures on rearing salmonids across an extensive region.

Optimizing performance efficiency of algal-bacterial-based wastewater treatment system using response surface methodology

Developing wastewater treatment technologies is crucial for enhancing treatment efficiency and promoting the reuse of treated water. In this article, the Algal-Bacterial integrated system (ABIS) is employed to treat municipal wastewater collected from the Zenin wastewater treatment plant in Giza governorate, Egypt. This system relies on the synergy between heterotrophic bacteria and microalgae to form a biofilm capable of absorbing contaminants from the wastewater. The main objective of this study is to optimize the factors influencing the treatment efficiency using response surface methodology derived from the experimental design software of Design Expert 6.0.8. The optimal results revealed that using an absorber amount of 0.18 g/L for a hydraulic retention time of 1.4 d provides removal efficiency of BOD, COD, TSS, and turbidity are 88.57%, 82.7 %, 94.90%, and 95.91%, respectively. These findings were experimentally and statistically verified with an accuracy exceeding a 95% confidence interval and 95% prediction interval. The physicochemical characteristics, algal community structure, as well as the density of total coliform, fecal coliform, and Escherichia coli (E. coli), were determined for the treated wastewater and compared to the Egyptian code of practice for the use of treated municipal wastewater for agricultural purposes to assess its suitability for unrestricted irrigation.

Pairwise Comparison of Heavy Dynamic Strength and Fast Dynamic Strength Interventions on Sprint Performance: A Systematic Review and Meta-Analysis.

Pairwise comparison of heavy dynamic strength and fast dynamic strength interventions on sprint performance: a systematic review and meta-analysis. J Strength Cond Res 38(9): 1675-1686, 2024-Previous studies have shown that both heavy dynamic strength (HDS) and fast dynamic strength (FDS) training can be used to improve sprint performance; however, a review and meta-analysis investigating pairwise studies that compare these two training interventions have not been performed. The aims of the study were to systematically review and analyze HDS and FDS training methodologies and evaluate their effect size difference, in pairwise comparison studies to determine and compare their effects on sprint performance. Databases were systematically searched using Boolean phrasing to identify eligible articles, and meta-analyses were performed on the extracted data. Seven studies met the inclusion criteria, which resulted in data from 138 subjects across 24 separate sprint assessments. Overall, there was a small effect in favor of FDS (standardized mean difference = 0.27, 95% confidence intervals [-0.07; 0.60], 95% prediction intervals [-1.01; 1.55]), but this was deemed not significant because of the wide-ranging prediction intervals. There is no distinguishable difference between HDS and FDS training on sprint performance. The wide-ranging prediction intervals suggest the variability is too great to determine whether one training type is more effective than the other. Practitioners should consider the individual needs of their athletes when deciding which training type to use for long-term sprint development.

An improved interval prediction method for recurrence period wind speed

Based on the improved interval operation theory, an improved expression of the return period wind speed interval prediction is constructed by using an approximate first-order Taylor series expansion. According to the measured wind speed data in Beijing, Jinan, Nanjing, Wuxi, Shanghai and Shenzhen, the improved method and the traditional method are respectively used to predict the interval of the return period wind speed. Furthermore, the interval results predicted by the improved method and the traditional method are compared and analyzed under the same confidence level. Results show that the improved method has good applicability for different parameter estimation methods under the condition of certain extreme value distribution model, and the interval prediction results of the return period wind speed are basically stable. Compared with the interval results predicted by the traditional method, the interval predicted by the improved method is more likely to be close to or contain the exact solution of the return period wind speed, which has higher prediction accuracy. In addition, the calculation process of the improved method is relatively simple and can realize the simplified calculation of interval prediction.

Ensemble Conformalized Quantile Regression for Probabilistic Time Series Forecasting.

This article presents a novel probabilistic forecasting method called ensemble conformalized quantile regression (EnCQR). EnCQR constructs distribution-free and approximately marginally valid prediction intervals (PIs), which are suitable for nonstationary and heteroscedastic time series data. EnCQR can be applied on top of a generic forecasting model, including deep learning architectures. EnCQR exploits a bootstrap ensemble estimator, which enables the use of conformal predictors for time series by removing the requirement of data exchangeability. The ensemble learners are implemented as generic machine learning algorithms performing quantile regression (QR), which allow the length of the PIs to adapt to local variability in the data. In the experiments, we predict time series characterized by a different amount of heteroscedasticity. The results demonstrate that EnCQR outperforms models based only on QR or conformal prediction (CP), and it provides sharper, more informative, and valid PIs.

Short-term wind power forecasting based on multi-scale receptive field-mixer and conditional mixture copula

Integrating renewable wind power (WP) into the grid exacerbates variability and challenges reliability. Establishing an effective forecasting system is crucial for risk avoidance, but achieving effective, interpretable, and accurate predictions remains an obstacle due to the high volatility and uncertainty of WP. This study proposes an innovative and trustworthy wind power forecasting system combined with point and interval predictions. Specifically, the adaptive multi-scale convolutional receptive field (AMSCF) combined with gated recurrent unit (GRU) for WP point prediction is proposed to capture the spatiotemporal dependencies of WP generation. This AMSCF method uses different size receptive fields, fuses them using channel dimension concatenation, operates pooling layer and adaptive fusion through field attention, which extracts hierarchical features for WP forecasting. Then conditional mixture Copula (Con-mCopula) function integrated with multiobjective optimizer is established for WP interval prediction, which reduces the WP interval prediction accuracy error as some information neglected. The wind power cluster dataset from Germany was selected for experiments. Based on the results, for both datasets, the designed prediction system can achieve better point and interval prediction performance resulting in a reduction of mean absolute error (MAE) by approximately 40 % compared to other existing components in the market. This study deconstructs the “Black Box” in deep network for energy forecasting through interpretability analysis, which accelerates the development of AI technology in WP forecasting system and provides decision-makers with trustworthy WP forecasting assistance.

Constructing Interval Forecasts for Solar and Wind Energy Using Quantile Regression, ARCH and Exponential Smoothing Methods

The research reported in this article focuses on a comparison of two different approaches to setting error bounds, or prediction intervals, on short-term forecasting of solar irradiation as well as solar and wind farm output. Short term in this instance relates to the time scales applicable in the Australian National Electricity Market (NEM), which operates on a five-minute basis throughout the year. The Australian Energy Market Operator (AEMO) has decided in recent years that, as well as point forecasts of energy, it is advantageous for planning purposes to have error bounds on those forecasts. We use quantile regression as one of the techniques to construct the bounds. This procedure is compared to a method of forecasting the conditional variance by use of either ARCH/GARCH or exponential smoothing, whichever is more appropriate for the specific application. The noise terms for these techniques must undergo a normalising transformation before their application. It seems that, for certain applications, quantile regression performs better, and the other technique for some other applications.

Bronchial Valves for Persistent Air Leak: A Systematic Review and Meta-analysis.

Patients with persistent air leak (PAL) pose a therapeutic challenge to physicians, with prolonged hospital stays and high morbidity. There is little evidence on the efficacy and safety of bronchial valves (BV) for PAL. We systematically searched the PubMed and Embase databases to identify studies evaluating the efficacy and safety of BV for PAL. We calculated the success rate (complete resolution of air leak or removal of intercostal chest drain after bronchial valve placement and requiring no further procedures) of BV for PAL in individual studies. We pooled the data using a random-effects model and examined the factors influencing the success rate using multivariable meta-regression. We analyzed 28 observational studies (2472 participants). The pooled success rate of bronchial valves in PAL was 82% (95% confidence intervals, 75 to 88; 95% prediction intervals, 64 to 92). We found a higher success rate in studies using intrabronchial valves versus endobronchial valves (84% vs. 72%) and in studies with more than 50 subjects (93% vs. 77%). However, none of the factors influenced the success rate of multivariable meta-regression. The overall complication rate was 9.1% (48/527). Granulation tissue was the most common complication reported followed by valve migration or expectoration and hypoxemia. Bronchial valves are an effective and safe option for treating PAL. However, the analysis is limited by the availability of only observational data.

Common double-lumen tube selection methods overestimate adequate tube sizes in individual patients – a 3D reconstruction study

BackgroundAppropriate selection of double-lumen tube sizes for one-lung ventilation is crucial to prevent airway damage. Current selection methods rely on demographic factors or 2D radiography. Prediction of left bronchial diameter is indispensable for choosing the adequate tube size. This prospective observational study investigates if current selection methods sufficiently predict individuals’ left bronchial diameters for DLT selection compared to the 3D reconstruction.Methods100 patients necessitating thoracic surgery with one-lung ventilation and left-sided double-lumen tubes, ≥ 18 years of age, and a set of chest X-rays and 2D thorax CT scans for 3D reconstruction of the left main bronchus were included between 07/2021 and 06/2023. The cross-validated prediction error and the width of the 95%-prediction intervals of the 3D left main bronchial diameter utilizing linear prediction models were based on current selection methods.ResultsThe mean bronchial diameter in 3D reconstruction was 13.6 ± 2.1 mm. The ranges of the 95%-prediction intervals for the bronchial diameter were 6.4 mm for demographic variables, 8.3 mm for the tracheal diameter from the X-ray, and 5.9 mm for bronchial diameter from the 2D-CT scans. Current methods violated the suggested ‘≥1 mm’ safety criterion in up to 7% (men) and 42% (women). Particularly, 2D radiography overestimated women’s left bronchial diameter. Current methods even allowed the selection of double-lumen tubes with bronchial tube sections greater than the bronchial diameter in women.ConclusionsNeither demographic nor 2D-radiographic methods sufficiently account for the variability of the bronchial diameter. Wide 95%-prediction intervals for the bronchial diameter hamper accurate individual double-lumen tube selection. This increases women’s risk of bronchial damage, particularly if they have other predisposing factors. These patients may benefit from 3D reconstruction of the left main bronchus.Trial registrationNot applicable.

A decade of curtailment studies demonstrates a consistent and effective strategy to reduce bat fatalities at wind turbines in North America

Abstract There is a rapid, global push for wind energy installation. However, large numbers of bats are killed by turbines each year, raising concerns about the impacts of wind energy expansion on bat populations. Preventing turbine blades from spinning at low wind speeds, referred to as curtailment, is a method to reduce bat fatalities, but drawing consistent inference across studies has been challenging. We compiled publicly available studies that evaluated curtailment at six wind energy facilities in North America across 10 years. We used meta‐regression of 29 implemented treatments to determine fatality reduction efficacy as well as sources of variation influencing efficacy. We also estimated species‐specific fatality reduction for three species that comprise most fatalities in North America: hoary bat (Lasiurus cinereus), eastern red bat (Lasiurus borealis) and silver‐haired bat (Lasionycteris noctivagans). We found that curtailment reduced total bat fatalities by 33% with every 1.0 ms−1 increase in curtailment wind speed. Estimates of the efficacy for the three target species were similar (hoary bats: 28% per ms−1, 95% CI: 0.4%–48%, eastern red bats: 32% per ms−1, 95% CI: 13%–47% and silver‐haired bats: 32% per ms−1, 95% CI: 3%–53%). Across multiple facilities and years, a 5.0 ms−1 cut‐in speed was estimated to reduce total bat fatalities by an average of 62% (95% CI: 54%–69%). Mortality reductions at individual facilities in any given year were estimated to fall between 33%–79% (95% prediction interval). Inter‐annual differences rather than inter‐site or turbine characteristics accounted for most of the variation in efficacy rates. Species‐specific average mortality reduction at 5.0 ms−1 curtailment wind speed was 48% (95% CI: 24%–64%) for hoary bats, 61% (95% CI: 42%–74%) for eastern red bats and 52% (95% CI: 30%–66%) for silver‐haired bats. Practical implication. curtailment reduced bat mortality at wind turbines in this North American study. Efficacy increased proportionally as curtailment speed is raised, and patterns and rates of efficacy were similar across species. This indicates that curtailment is an effective strategy to reduce bat fatalities at wind energy facilities, but exploration of further refinements could both minimize bat mortality and maximize energy production.

Short-term wind farm cluster power point-interval prediction based on graph spatio-temporal features and S-Stacking combined reconstruction

Wind energy is becoming increasingly competitive, Accurate and reliable multi-engine wind power forecasts can reduce power system operating costs and improve wind power consumption capacity. Existing research on wind power forecasting has neglected the importance of interval forecasting using clusters of wind farms to capture spatial characteristics and the objective selection of forecasting sub-learners, leading to increased uncertainty and risk in system operation. This paper proposes a new “decomposition-aggregation-multi-model parallel prediction” method. The data set is pre-processed by a decomposition-aggregation strategy and spatial feature extraction, and then a Stacking model with multiple parallel sub-learners selected by bootstrap method is used for point and interval forecasting. Experiments and discussions are conducted based on 15-min resolution wind power data from a cluster dataset of a wind farm in northwest China. The experimental results indicate that the method achieves higher accuracy and reliability in both point prediction and interval prediction than other comparative models, with a root mean square error value of 7.47 and an average F value of 1.572, which can provide a reliable reference for power generation planning from wind farm clusters.

Follow-up of Cystic Pineal Glands in Retinoblastoma Patients Does Not Increase Detection of Pineal Trilateral Retinoblastoma

PurposeTo evaluate the effectiveness of baseline screening and follow-up with MRI to detecting trilateral retinoblastoma (TRb) and assessing the risk of TRb development. DesignProspective multicenter cohort study MethodsA total of 607 retinoblastoma patients from 2012 through 2022 were included and followed up until 1-9-2023. At each center a neuroradiologist categorized pineal glands on baseline and follow-up scans into four groups: (A) normal, (B) cystic gland, (C) suspicious gland, or (D) TRb. Different follow-up schedules were assigned to each category. Categories (B) and (C) were followed-up with MRI after approximately 3-months and after another 3 months if suspicion remained. On each MRI, they measured the height and width, evaluated the aspect (solid, partly cystic and completely cystic) of the pineal gland and evaluated radiological features suspicious of pineal TRb. The effectiveness of the current TRb screening method was assessed by evaluating its sensitivity and specificity to detect TRb. Determining the TRb incidence was a secondary outcome measure. ResultsHeritable retinoblastoma patients had a risk of 3.78% to develop TRb. One out of four pineal TRbs was detected during a follow-up scan and four out of five non-pineal TRbs were detected on the baseline MRI. Screening for pineal TRb had a sensitivity of 25% and specificity of 100%, for non-pineal TRb the sensitivity was 80%. It required 494 follow-up scans to detect one pineal TRb. However, when restricting the follow-up to solely suspicious glands, only 22 scans were required to detect one pineal TRb. ConclusionDuring extended follow-up after baseline MRI, only one pineal trilateral retinoblastoma was detected in our study. Follow-up after three months should be restricted to patients with a suspicious pineal gland defined as irregularly thickening of the cyst wall (>2mm), fine nodular aspect of the cyst wall or when a solid or cystic gland exceeds the upper 99% prediction interval for size; patients with an unsuspicious cystic gland should not be followed up. Baseline MRI screening was able to detect most non-pineal trilateral retinoblastomas.

Evaluating the forecasting performance of ensemble sub-epidemic frameworks and other time series models for the 2022-2023 mpox epidemic.

During the 2022-2023 unprecedented mpox epidemic, near real-time short-term forecasts of the epidemic's trajectory were essential in intervention implementation and guiding policy. However, as case levels have significantly decreased, evaluating model performance is vital to advancing the field of epidemic forecasting. Using laboratory-confirmed mpox case data from the Centers for Disease Control and Prevention and Our World in Data teams, we generated retrospective sequential weekly forecasts for Brazil, Canada, France, Germany, Spain, the United Kingdom, the United States and at the global scale using an auto-regressive integrated moving average (ARIMA) model, generalized additive model, simple linear regression, Facebook's Prophet model, as well as the sub-epidemic wave and n-sub-epidemic modelling frameworks. We assessed forecast performance using average mean squared error, mean absolute error, weighted interval scores, 95% prediction interval coverage, skill scores and Winkler scores. Overall, the n-sub-epidemic modelling framework outcompeted other models across most locations and forecasting horizons, with the unweighted ensemble model performing best most frequently. The n-sub-epidemic and spatial-wave frameworks considerably improved in average forecasting performance relative to the ARIMA model (greater than 10%) for all performance metrics. Findings further support sub-epidemic frameworks for short-term forecasting epidemics of emerging and re-emerging infectious diseases.

A Stochastic Recurrent Encoder Decoder Network for Multistep Probabilistic Wind Power Predictions.

In this article, a stochastic recurrent encoder decoder neural network (SREDNN), which considers latent random variables in its recurrent structures, is developed for the first time for the generative multistep probabilistic wind power predictions (MPWPPs). The SREDNN enables the stochastic recurrent model under the encoder-decoder framework to engage exogenous covariates to produce better MPWPP. The SREDNN consists of five components, the prior network, the inference network, the generative network, the encoder recurrent network, and the decoder recurrent network. The SREDNN is equipped with two critical advantages compared with conventional RNN-based methods. First, the integration over the latent random variable builds an infinite Gaussian mixture model (IGMM) as the observation model, which drastically increases the expressiveness of the wind power distribution. Secondly, hidden states of the SREDNN are updated in a stochastic way, which builds an infinite mixture of the IGMM for describing the ultimate wind power distribution and enables the SREDNN to model complex patterns across wind speed and wind power sequences. Computational experiments are conducted on a dataset of a commercial wind farm having 25 wind turbines (WTs) and two publicly assessable WT datasets to verify the advantages and effectiveness of the SREDNN for MPWPP. Experimental results show that the SREDNN achieves a lower negative form of the continuously ranked probability score (CRPS*) as well as a superior sharpness and comparable reliability of prediction intervals by comparing against considered benchmarking models. Results also show the clear benefit gained from considering latent random variables in SREDNN.

Modeling of anti-spike IgG and neutralizing antibody waning after anti-SARS-CoV-2 mRNA vaccination

At present, mRNA-based vaccines against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) are being administered on a global scale. While the efficacy of mRNA vaccines has been demonstrated, several unknowns remains. For example, as the number of booster vaccinations increases, there are uncertainties regarding how long effects of a vaccine will last and how much individual variability exists. In this study, to predict the duration of vaccine efficacy, we modeled the kinetics of antibody levels for each SARS-CoV-2 vaccination dose, incorporating predictive intervals to estimate the duration of vaccine efficacy and to account for variability among individuals. A total of 3,059 serum samples from 1,346 participants were assayed to quantify IgG antibodies specific for the S1 subunit of the S protein (anti-S1 IgG) and neutralizing antibody activities against SARS-CoV-2. A power law model was used to simulate the decay of antibody titers following vaccination, and models were constructed to assess antibody level kinetics after the second, third, fourth, and fifth vaccinations. The models assumed that booster vaccinations would significantly reduce the decline in anti-S antibody and neutralizing antibody levels, resulting in levels being maintained for a longer period. No significant differences in the decay rate of antibody levels were observed among age groups, yet the peak titers of antibody levels were significantly higher in the ≤ 39 age group than in the ≥ 60 age group following the second vaccination; these differences were not observed after the third and fourth vaccinations. The modeling of antibody level kinetics after vaccination is considered to be useful for understanding the immune status of mRNA vaccine recipients.

Ultra-short-term wind power forecasting method based on multi-variable joint extraction of spatial-temporal features

Accurate and reliable wind power forecasting is imperative for wind power stations' stable and efficient operation. Information such as wind speed and wind direction in the same wind field has spatial-temporal differences. Considering the spatial-temporal changes in wind fields can improve model prediction accuracy. However, existing methods suffer from limited ability to capture correlation features among variables, information loss in spatial-temporal feature extraction, and neglect short-term temporal features. This paper introduces a novel ultra-short-term wind power forecasting method based on the combination of a deep separable convolutional neural network (DSCNN) and long- and short-term time-series network (LSTNet), incorporating maximum information coefficient (MIC) to realize multi-variable joint extraction of spatial-temporal features. The method utilizes MIC to jointly analyze and process the multi-variate variables before spatial-temporal feature extraction to avoid information redundancy. The spatial features between input variables and wind power are extracted by deep convolution and pointwise convolution in DSCNN. Then, a convolutional neural network and gated recurrent unit in LSTNet are combined to capture long-term and short-term temporal features. In addition, an autoregressive module is employed to accept features extracted by MIC to enhance the model's learning of temporal features. Based on real datasets, the performance of models is validated through comprehensive evaluation experiments such as comparison experiments, ablation experiments, and interval prediction methods. The results show that the proposed method reduces mean absolute error by up to 4.66% and provides more accurate prediction intervals, verifying the accuracy and effectiveness of the proposed method.