Search

This paper proposes a novel approach to explain the predictions made by data-driven methods. Since such predictions rely heavily on the data used for training, explanations that convey information about how the training data affects the predictions are useful. The paper proposes a novel approach to quantify how different data-clusters of the training data affect a prediction. The quantification is based on Shapley values, a concept which originates from coalitional game theory, developed to fairly distribute the payout among a set of cooperating players. A player's Shapley value is a measure of that player's contribution. Shapley values are often used to quantify feature importance, ie. how features affect a prediction. This paper extends this to cluster importance, letting clusters of the training data act as players in a game where the predictions are the payouts. The novel methodology proposed in this paper lets us explore and investigate how different clusters of the training data affect the predictions made by any black-box model, allowing new aspects of the reasoning and inner workings of a prediction model to be conveyed to the users. The methodology is fundamentally different from existing explanation methods, providing insight which would not be available otherwise, and should complement existing explanation methods, including explanations based on feature importance. [ABSTRACT FROM AUTHOR]

Predicting the rate of penetration (ROP) is important for optimizing drilling parameters, improving drilling efficiency, and optimizing economic benefits throughout the drilling process. The current prediction model of ROP based on machine learning algorithms does not consider the interference of outliers. Therefore, in this study, we propose a method to predict ROP based on fuzzy support vector regression (FSVR). First, appropriate input parameters were selected from the controllable parameters. Second, based on the local outlier factor, a fuzzy membership degree was assigned to each sample. Finally, the sample with the fuzzy membership value was input into the model for ROP prediction. The results demonstrated that the goodness of fit (R2) of the improved FSVR model is 0.9634, and the mean absolute error is 0.1974. Compared with standard SVR and other models, the improved FSVR model has a stronger anti-interference ability, smaller prediction error for normal samples, and higher accuracy. [ABSTRACT FROM AUTHOR]

In response to the current situation of poor adaptive learning performance in Agent production and sales negotiation and dynamic changes in negotiation environment, the author proposes a method based on machine learning algorithms. Consider the impact of conflict level, cooperation possibility, and negotiation remaining time on negotiations in a dynamic negotiation environment, and use the entropy method to determine the weights of three influencing factors and perform linear weighting. Based on the differences in current negotiation topics, a concession amplitude prediction model based on dynamic selective ensemble learning is constructed, and an optimization strategy for supply chain production and sales negotiation is proposed. The experimental results indicate that, in the adaptive negotiation strategy of a regular SVM single learning machine, the joint utility of the most successfully negotiated agents falls within the interval [0.55, 0.70], while the author's ensemble learning strategy mainly focuses on [0.6, 0.8], the author's strategy is relatively superior to ordinary learning strategies in terms of both the number of successfully negotiated agents and the joint utility. Compared with the single learning machine negotiation strategy, this strategy improves the success rate and joint utility of Agent adaptive learning, and ensures the benefits of both production and sales in the supply chain, achieving a mutually beneficial situation for both parties in cooperation. [ABSTRACT FROM AUTHOR]

Background and objective: Surgical site infection (SSI) after dermatological surgery is associated with poor outcomes. Developing clinical prediction rules based on the predicted probability of infection might encourage guided prophylaxis and judicious prescribing. The purpose of this study was to develop a clinical prediction rule based on identified risk factors for SSI in a large general practice patient cohort. Methods: We examined a large, pooled dataset from four randomised controlled trials performed in a regional centre of North Queensland, Australia. Multivariable logistic regression identified a prediction model. Bootstrapping was used for internal validation. A scoring system was based on predicted probabilities of infection. Results: The final prediction rule included age >55 years and the anatomical site, histology and complexity of the excision. The area under the curve was 0.704. Discussion: Our prediction rule encourages judicious use of prophylaxis in clinical practice. [ABSTRACT FROM AUTHOR]

In the realm of credit card fraud detection, the landscape is continually evolving, demanding innovative approaches to stay ahead of increasingly sophisticated fraudulent activities. Our research pioneers a groundbreaking methodology that amalgamates the power of bipartite graph visualization with advanced machine learning techniques. This fusion yields a comprehensive framework capable of effectively evaluating the efficacy of a random forest classifier in uncovering fraudulent credit card transactions. Our study showcases the compelling application of this methodology, offering a paradigm shift in how we analyze and understand credit card fraud detection systems. By seamlessly integrating machine learning algorithms with network analysis, we provide a holistic view of the data, unveiling intricate patterns hidden within. At the heart of our approach lies the innovative use of bipartite graphs, which serve as a dynamic visual bridge between model predictions and real-world outcomes. This visual representation not only enhances interpretability but also facilitates a deeper understanding of the classifier’s performance. By visually mapping the relationships between transactions and their respective classifications, our methodology offers actionable insights into both successful detection and potential areas for improvement. Empowering analysts and stakeholders, our approach facilitates informed decisionmaking by enabling them to fine-tune model parameters and enhance the overall effectiveness of fraud detection systems. Through this synergy between cutting-edge machine learning and network analysis techniques, we provide a powerful tool to combat the critical challenge of credit card fraud prevention. Step into the future of fraud detection with our innovative methodology, where every transaction is scrutinized with precision, and where security is not just a possibility, but a promise fulfilled. [ABSTRACT FROM AUTHOR]

Wireless sensor networks (WSNs) are a pivotal technology in the modern era, enabling the monitoring and sensing of environmental conditions across vast areas with unprecedented precision and flexibility. At the heart of WSNs lie crucial challenges such as optimizing coverage, extending network lifetime, and strategizing node deployment to ensure efficient operation while conserving energy. This paper introduces the coverage-aware optimized sensor activation and deployment (CAOSAD) Strategy, a novel methodology designed to address these challenges. By integrating advanced node placement algorithms and scheduling techniques, the EcoNet lifespan maximization (ELM) strategy significantly enhances area and target coverage, minimizes energy consumption, and thereby prolongs the network’s operational lifespan. We present a comprehensive framework that dynamically adjusts node activity based on a predictive model, ensuring robust coverage and connectivity with minimal energy expenditure. Through a series of simulations, the ELM strategy demonstrates a substantial improvement in network sustainability compared to existing methodologies, offering a promising approach for the development of future WSNs. By focusing on the synergy between coverage optimization, energy-efficient node deployment, and innovative scheduling algorithms, this paper contributes a ground-breaking perspective to the research and application of WSNs, setting a new benchmark for the design of eco-friendly and durable sensing infrastructures. [ABSTRACT FROM AUTHOR]

In this paper, an integrated tube robust iterative learning model predictive control (Tube‐RILMPC) strategy based on the dynamic partial least squares (DyPLS) identified algorithm is proposed. The problems of large amount of online data calculation and input and output variable dimensions disaster in the original variable space are solved. This integrated Tube‐RILMPC strategy enhances the tracking performance and robustness of two‐dimensional (2D) control system. The output trajectories of system are located in the tube the nominal trajectory, which reduces the modelling error caused by the uncertain model. Based on the worst‐case performance index of ellipsoidal uncertainty and polytopic uncertainty, a robust iterative learning control (ILC) strategy is designed. Finally, the superiority of the proposed control algorithm is verified by comparative simulation. [ABSTRACT FROM AUTHOR]

This paper presents an advanced control strategy for a continuous biodiesel production plant based on a steady‐state optimizer and model predictive control (MPC). The proposed control system aims to optimize the production process and maintain product quality within required specifications. First, two steady‐state optimizers were developed with the aim of minimizing the steady‐state deviations of the manipulated and controlled variables and minimizing the biodiesel production cost. An MPC was then formulated to track the set points imposed by the steady‐state optimizers in real time and manipulate the control inputs accordingly. The scope of this work is limited to measured disturbances only. The effectiveness of the proposed control strategy is demonstrated through dynamic simulation studies performed using HYSYS and MATLAB. The results obtained using the proposed control methodology show significant improvements in performance compared to conventional control strategies. Furthermore, it avoids the quality problem reflected in the amount of water in the final product that the original plant presented due to an inadequate design of the control strategy. Overall, the results of this research indicate that the proposed advanced control strategy has the potential to improve the efficiency and profitability of continuous biodiesel production plants. [ABSTRACT FROM AUTHOR]

Experiments suggest that people fail to take into account interdependencies between their choices—they do not broadly bracket. Researchers often instead assume people narrowly bracket, but existing designs do not test it. We design a novel experiment and revealed preference tests for how someone brackets their choices. In portfolio allocation under risk, social allocation, and induced-value shopping experiments, 40–43 percent of subjects are consistent with narrow bracketing, and 0–16 percent with broad bracketing. Adjusting for each model's predictive precision, 74 percent of subjects are best described by narrow bracketing, 13 percent by broad bracketing, and 6 percent by intermediate cases. (JEL D12, D81, D91) [ABSTRACT FROM AUTHOR]

Background: High lactate to albumin ratio (LAR) has been reported to be associated to with poor prognosis in patients admitted to the intensive care unit (ICU). However, its role in predicting in-hospital mortality in AF patients admitted to ICU has not been explored. Methods: The Medical Information Mart for Intensive Care-IV (MIMIC-IV) database was used to retrieve information on patients who had been diagnosed with AF. X-tile software was utilized to determine the optimal cut-off LAR. Area under the receiver operating characteristic curves (AUC), calibration plots, and decision curve analysis (DCA) were conducted to assess the prediction performance of LAR for in-hospital mortality. Results: Finally, 8,287 AF patients were included and 1,543 death (18.6%) occurred. The optimal cut-off value of LAR is 0.5. Patients in lower LAR (< 0.5) group showed a better in-hospital survival compared to patients in higher LAR (≥ 0.5) group (HR: 2.67, 95%CI:2.39–2.97, P < 0.001). A nomogram for in-hospital mortality in patients with AF was constructed based on multivariate Cox analysis including age, CCI, β blockers usage, APSIII, hemoglobin and LAR. This nomogram exhibited excellent discrimination and calibration abilities in predicting in-hospital mortality for critically ill AF patients. Conclusion: LAR, as a readily available biomarker, can predict in-hospital mortality in AF patients admitted to the ICU. The nomogram that combined LAR with other relevant variables performed exceptionally well in terms of predicting in-hospital mortality. [ABSTRACT FROM AUTHOR]

Obesity is an abnormal and potentially dangerous condition caused by excess body fat accumulation. The number of people with obesity is increasing worldwide. Obesity is the primary cause of various diseases; therefore, it is crucial to make efforts to control body weight. Identifying the factors that influence men with obesity to attempt to control and not control their weight is essential. The objective of this study was to create a prediction model for weight control experience among Korean men in their 30 s and 40 s. We analyzed data from the 2022 Community Health Survey and included 12,311 men who were overweight or obese. The men were divided into two groups based on their weight control experience: (1) Yes group (n = 9405) and (2) No group (n = 2906). Chi-square and independent t-tests were used to compare general and health-related characteristics between the groups. Decision tree analysis was used to build a prediction model for weight control experience. A split-sample test was conducted to validate the model. From the results of this study, various models predicting weight control experience were derived. From the decision tree model without setting the first node, those who weighed below average, had a high school diploma or less, and did not know their blood sugar levels had the highest probability of not controlling their weight at 55.3%. In the prediction model where the first node was set to age, those in their 40 s who thought their weight was below average and were unaware of their blood sugar levels had the highest rate of not trying to control their weight at 50.1%. In the prediction model where the first node was set to BMI, those who were overweight but thought their weight was below average and had a high school diploma or less had the highest rate of not trying to control their weight at 51.5%. There is an urgent need to provide obesity prevention and management education to those who have no weight control experience, particularly those at high risk, as identified in this study. [ABSTRACT FROM AUTHOR]

Background: Multiple organ failure/dysfunction syndrome (MOF/MODS) is a major cause of mortality and morbidity among severe trauma patients. Current clinical practices entail monitoring physiological measurements and applying clinical score systems to diagnose its onset. Instead, we aimed to develop an early prediction model for MOF outcome evaluated soon after traumatic injury by performing machine learning analysis of genome-wide transcriptome data from blood samples drawn within 24 h of traumatic injury. We then compared its performance to baseline injury severity scores and detection of infections. Methods: Buffy coat transcriptome and linked clinical datasets from blunt trauma patients from the Inflammation and the Host Response to Injury Study ("Glue Grant") multi-center cohort were used. According to the inclusion/exclusion criteria, 141 adult (age ≥ 16 years old) blunt trauma patients (excluding penetrating) with early buffy coat (≤ 24 h since trauma injury) samples were analyzed, with 58 MOF-cases and 83 non-cases. We applied the Least Absolute Shrinkage and Selection Operator (LASSO) and eXtreme Gradient Boosting (XGBoost) algorithms to select features and develop models for MOF early outcome prediction. Results: The LASSO model included 18 transcripts (AUROC [95% CI]: 0.938 [0.890–0.987] (training) and 0.833 [0.699–0.967] (test)), and the XGBoost model included 41 transcripts (0.999 [0.997–1.000] (training) and 0.907 [0.816–0.998] (test)). There were 16 overlapping transcripts comparing the two panels (0.935 [0.884–0.985] (training) and 0.836 [0.703–0.968] (test)). The biomarker models notably outperformed models based on injury severity scores and sex, which we found to be significantly associated with MOF (APACHEII + sex—0.649 [0.537–0.762] (training) and 0.493 [0.301–0.685] (test); ISS + sex—0.630 [0.516–0.744] (training) and 0.482 [0.293–0.670] (test); NISS + sex—0.651 [0.540–0.763] (training) and 0.525 [0.335–0.714] (test)). Conclusions: The accurate assessment of MOF from blood samples immediately after trauma is expected to aid in improving clinical decision-making and may contribute to reduced morbidity, mortality and healthcare costs. Moreover, understanding the molecular mechanisms involving the transcripts identified as important for MOF prediction may eventually aid in developing novel interventions. [ABSTRACT FROM AUTHOR]

Background: To investigate how successfully the classification of patients with and without dental anomalies was achieved through four experiments involving different dental anomalies. Methods: Lateral cephalometric radiographs (LCRs) from 526 individuals aged between 14 and 22 years were included. Four experiments involving different dental anomalies were created. Experiment 1 included the total dental anomaly group and control group (CG). Experiment 2 only had dental agenesis and a CG. Experiment 3 consisted of only palatally impacted canines and the CG. Experiment 4 comprised patients with various dental defects (transposition, hypodontia, agenesis-palatally affected canine, peg-shaped laterally, hyperdontia) and the CG. Twelve sella measurements and assessments of the ponticulus posticus and posterior arch deficiency were given as input. The target was to distinguish between anomalies and controls. The CatBoost algorithm was applied to classify patients with and without dental anomalies. Results: In order from lowest to highest, the predictive accuracies of the experiments were as follows: experiment 4 < experiment 2 < experiment 3 < experiment 1. The sella area (SA) (mm2) was the most important variable in experiment 1. The most significant variable in prediction model of experiment 2 was sella height posterior (SHP) (mm). Sella area (SA) (mm2) was again the most relevant variable in experiment 3. The most important variable in experiment 4 was sella height median (SHM) (mm). Conclusions: Every prediction model from the four experiments prioritized different variables. These findings may suggest that related research should focus on specific traits from a diagnostic perspective. [ABSTRACT FROM AUTHOR]

We evaluate the skill and sources of skill in initialized seasonal climate forecasts of monthly global mean temperature from the North American Multi‐Model Ensemble (NMME) during the period 1991–2024. The forecasts demonstrate skill in addition to that from the long‐term trend, and that skill is primarily attributable to ENSO. However, the skill varies seasonally, with skill being lowest for target periods during Northern Hemisphere summer. Single model ensembles show underdispersion at short leads, while the multi‐model ensemble is overdispersed, suggesting initial condition errors and highlighting the importance of model initialization for quantification of forecast uncertainty. Lead‐time dependent errors in global mean temperature trends appear related to Pacific trend errors. The multi‐model mean captured the overall trend but underestimated the record‐breaking temperatures of 2023. Forecasts for the remainder of 2024 indicate cooling by the end of the year. Plain Language Summary: Our study looked at how well current climate forecast models can predict the global average temperature up to a year in advance. We found that these models can predict temperature changes better than just looking at long‐term trends, and that their ability to do so was related to the climate phenomenon called ENSO (El Niño‐Southern Oscillation). However, the accuracy of these predictions depends on the time of year, being less accurate when predicting temperatures during the Northern Hemisphere's summer months. We also found that individual models often have too narrow a range of predictions, while combining multiple models results in a range that is too wide. Despite these deficiencies, the combined model predictions generally followed the observed temperatures but failed to predict the extreme high temperatures of 2023. Looking ahead, the models suggest a cooling trend by the end of 2024. This research improves our understanding of what current climate forecast models can tell us about global mean temperature in the short term and highlights areas where improvement is needed. Key Points: Initialized seasonal climate forecasts of global mean temperature show skill beyond the trend and that skill is largely related to ENSOForecast skill is lowest for target months during northern hemisphere summer due to forecast amplitudes that are too largeAt short leads single model ensembles are underdispersed and the multi‐model ensemble is overdispersed, which suggests initialization errors [ABSTRACT FROM AUTHOR]

Roughly one‐third of sudden stratospheric warming (SSW) events lack a strong canonical surface response, and this can lead to a forecast bust if a strong response was predicted. Hence, it is desirable to predict before SSW onset if an event will propagate downward. The predictability of the downward response of SSWs is considered in seven subseasonal‐to‐seasonal forecast models for 16 major SSWs between 1998 and 2022, a larger sample size than considered by previous works. The models successfully predict before SSW onset which SSWs have a stronger downward response to 100 hPa, however they struggle to predict which have a stronger tropospheric response. The downward response is stronger if the magnitude of the deceleration of the 10 hPa winds is more accurately predicted. Downward response is stronger for split and absorbing SSWs. In contrast, there is little relationship between SSWs whose onset can be predicted at earlier leads and the downward response. Plain Language Summary: The wintertime stratosphere typically features circumpolar strong westerly winds, but on occasion these strong winds can reverse and temperatures over the pole can rise by tens of degrees in an event known as a sudden stratospheric warming (SSW). Such an event increases the likelihood of extreme cold over Northern Eurasia and wet conditions in Southern Europe, however roughly a third of events do not feature such downward propagation. Sixteen SSW events have occurred in the Northern Hemisphere over the period 1998 to 2022, and this study considers whether the models that have contributed to the subseasonal to seasonal (S2S) database are able to distinguish which are downward propagating and which are not. We also explore the factors that govern downward propagation of the SSW signal in these models. Key Points: Downward versus non‐downward response of SSWs to 100 hPa is predicted by subseasonal prediction models ∼20 days before SSW onsetIn contrast, models struggle to predict the presence/absence of a tropospheric response before SSW onsetDownward response and predictability of SSWs unrelated; split and absorbing SSWs show stronger downward response in models and reanalysis [ABSTRACT FROM AUTHOR]

In this registered report, we propose to stress-test existing models for predicting the ideology-prejudice association, which varies in size and direction across target groups. Previous models of this relationship use the perceived ideology, status, and choice in group membership of target groups to predict the ideology-prejudice association across target groups. These analyses show that models using only the perceived ideology of the target group are more accurate and parsimonious in predicting the ideology-prejudice association than models using perceived status, choice, and all of the characteristics in a single model. Here, we stress-test the original models by testing the models' predictive utility with new measures of explicit prejudice, a comparative operationalization of prejudice, the Implicit Association Test, and additional target groups. In Study 1, we propose to directly test the previous models using an absolute measure of prejudice that closely resembles the measure used in the original study. This will tell us if the models replicate with distinct, yet conceptually similar measures of prejudice. In Study 2, we propose to develop new ideology-prejudice models for a comparative operationalization of prejudice using both explicit measures and the Implicit Association Test. We will then test these new models using data from the Ideology 2.0 project collected by Project Implicit. We do not have full access to this data yet, but upon acceptance of our Stage 1 registered report, we will gain access to the complete dataset. Currently, we have access to an exploratory subset of the data that we use to demonstrate the feasibility of the study, but its limited number of target groups prevents conclusions from being made. [ABSTRACT FROM AUTHOR]

We present a comparison of machine learning methods for the prediction of four quantitative traits in Arabidopsis thaliana. High prediction accuracies were achieved on individuals grown under standardized laboratory conditions from the 1001 Arabidopsis Genomes Project. An existing body of evidence suggests that linear models may be impeded by their inability to make use of non-additive effects to explain phenotypic variation at the population level. The results presented here use a nested cross-validation approach to confirm that some machine learning methods have the ability to statistically outperform linear prediction models, with the optimal model dependent on availability of training data and genetic architecture of the trait in question. Linear models were competitive in their performance as per previous work, though the neural network class of predictors was observed to be the most accurate and robust for traits with high heritability. The extent to which non-linear models exploit interaction effects will require further investigation of the causal pathways that lay behind their predictions. Future work utilizing more traits and larger sample sizes, combined with an improved understanding of their respective genetic architectures, may lead to improvements in prediction accuracy. [ABSTRACT FROM AUTHOR]

Objective: To identify variables associated with a patients' ability to reproducibly hold their breath for deep-inspiration breath-hold (DIBH) radiotherapy (RT) and to develop a predictive model for DIBH eligibility. Methods: This prospective, single-institution, IRB-approved observational study included women with left-sided breast cancer treated between January 2023 and March 2024. Patients underwent multiple breath-hold sessions over 2–3 consecutive days. DIBH waveform metrics and clinical factors were recorded and analysed. Logistic mixed modelling was used to predict DIBH eligibility, and a temporal validation cohort was used to assess model performance. Results: In total, 253 patients were included, with 206 in the model development cohort and 47 in the temporal validation cohort. The final logistic mixed model identified increasing average breath-hold duration (OR, 95% CI: 0.308, 0.104–0.910. p = 0.033) and lower amplitude (OR, 95% CI: 0.737, 0.641–0.848. p < 0.001) as significant predictors of DIBH eligibility. Increasing age was associated with higher odds of being ineligible for DIBH (OR, 95% CI: 1.040, 1.001–1.081. p = 0.044). The model demonstrated good discriminative performance in the validation cohort with an AUC of 80.9% (95% CI: 73.0-88.8). Conclusion: The identification of variables associated with DIBH eligibility and development of a predictive model has the potential to serve as a decision-support tool. Further external validation is required before its integration into routine clinical practice. [ABSTRACT FROM AUTHOR]

Background: Distributed statistical analyses provide a promising approach for privacy protection when analyzing data distributed over several databases. Instead of directly operating on data, the analyst receives anonymous summary statistics, which are combined into an aggregated result. Further, in discrimination model (prognosis, diagnosis, etc.) development, it is key to evaluate a trained model w.r.t. to its prognostic or predictive performance on new independent data. For binary classification, quantifying discrimination uses the receiver operating characteristics (ROC) and its area under the curve (AUC) as aggregation measure. We are interested to calculate both as well as basic indicators of calibration-in-the-large for a binary classification task using a distributed and privacy-preserving approach. Methods: We employ DataSHIELD as the technology to carry out distributed analyses, and we use a newly developed algorithm to validate the prediction score by conducting distributed and privacy-preserving ROC analysis. Calibration curves are constructed from mean values over sites. The determination of ROC and its AUC is based on a generalized linear model (GLM) approximation of the true ROC curve, the ROC-GLM, as well as on ideas of differential privacy (DP). DP adds noise (quantified by the ℓ 2 sensitivity Δ 2 (f ^) ) to the data and enables a global handling of placement numbers. The impact of DP parameters was studied by simulations. Results: In our simulation scenario, the true and distributed AUC measures differ by Δ AUC < 0.01 depending heavily on the choice of the differential privacy parameters. It is recommended to check the accuracy of the distributed AUC estimator in specific simulation scenarios along with a reasonable choice of DP parameters. Here, the accuracy of the distributed AUC estimator may be impaired by too much artificial noise added from DP. Conclusions: The applicability of our algorithms depends on the ℓ 2 sensitivity Δ 2 (f ^) of the underlying statistical/predictive model. The simulations carried out have shown that the approximation error is acceptable for the majority of simulated cases. For models with high Δ 2 (f ^) , the privacy parameters must be set accordingly higher to ensure sufficient privacy protection, which affects the approximation error. This work shows that complex measures, as the AUC, are applicable for validation in distributed setups while preserving an individual's privacy. [ABSTRACT FROM AUTHOR]

Background: Accurate prediction of subject recruitment, which is critical to the success of a study, remains an ongoing challenge. Previous prediction models often rely on parametric assumptions which are not always met or may be difficult to implement. We aim to develop a novel method that is less sensitive to model assumptions and relatively easy to implement. Methods: We create a weighted resampling-based approach to predict enrollment in year two based on recruitment data from year one of the completed GRIPS and PACE clinical trials. Different weight functions accounted for a range of potential enrollment trajectory patterns. Prediction accuracy was measured by Euclidean distance for enrollment sequence in year two, total enrollment over time, and total weeks to enroll a fixed number of subjects, against the actual year two enrollment data. We compare the performance of the proposed method with an existing Bayesian method. Results: Weighted resampling using GRIPS data resulted in closer prediction evidenced by better coverage of observed enrollment with the prediction intervals and smaller Euclidean distance from actual enrollment in year 2, especially when enrollment gaps were filled prior to the weighted resampling. These scenarios also produced more accurate predictions for total enrollment and number of weeks to enroll 50 participants. These same scenarios outperformed an existing Bayesian method for all 3 accuracy measures. In PACE data, using a reduced year 1 enrollment resulted in closer prediction evidenced by better coverage of observed enrollment with the prediction intervals and smaller Euclidean distance from actual enrollment in year 2, with the weighted resampling scenarios better reflecting the seasonal variation seen in year (1) The reduced enrollment scenarios resulted in closer prediction for total enrollment over 6 and 12 months into year (2) These same scenarios also outperformed an existing Bayesian method for relevant accuracy measures. Conclusion: The results demonstrate the feasibility and flexibility for a resampling-based, non-parametric approach for prediction of clinical trial recruitment with limited early enrollment data. Application to a wider setting and long-term prediction accuracy require further investigation. [ABSTRACT FROM AUTHOR]

Background: Determining brain atrophy is crucial for the diagnosis of neurodegenerative diseases. Despite detailed brain atrophy assessments using three-dimensional (3D) T1-weighted magnetic resonance imaging, their practical utility is limited by cost and time. This study introduces deep learning algorithms for quantifying brain atrophy using a more accessible two-dimensional (2D) T1, aiming to achieve cost-effective differentiation of dementia of the Alzheimer's type (DAT) from cognitively unimpaired (CU), while maintaining or exceeding the performance obtained with T1-3D individuals and to accurately predict ADspecific atrophy similarity and atrophic changes [W-scores and Brain Age Index (BAI)]. Methods: Involving 924 participants (478 CU and 446 DAT), our deep learning models were trained on cerebrospinal fluid (CSF) volumes from 2D T1 images and compared with 3D T1 images. The performance of the models in differentiating DAT from CU was assessed using receiver operating characteristic analysis. Pearson's correlation analyses were used to evaluate the relations between 3D T1 and 2D T1 measurements of cortical thickness and CSF volumes, AD-specific atrophy similarity, W-scores, and BAIs. Results: Our deep learning models demonstrated strong correlations between 2D and 3D T1-derived CSF volumes, with correlation coefficients r ranging from 0.805 to 0.971. The algorithms based on 2D T1 accurately distinguished DAT from CU with high accuracy (area under the curve values of 0.873), which were comparable to those of algorithms based on 3D T1. Algorithms based on 2D T1 image-derived CSF volumes showed high correlations in AD-specific atrophy similarity (r = 0.915), W-scores for brain atrophy (0.732 ≤ r ≤ 0.976), and BAIs (r = 0.821) compared with those based on 3D T1 images. Conclusion: Deep learning-based analysis of 2D T1 images is a feasible and accurate alternative for assessing brain atrophy, offering diagnostic precision comparable to that of 3D T1 imaging. This approach offers the advantage of the availability of T1-2D imaging, as well as reduced time and cost, while maintaining diagnostic precision comparable to T1-3D. [ABSTRACT FROM AUTHOR]

Introduction: The tumor microenvironment and IRGs are highly correlated with tumor occurrence, progression, and prognosis. However, their roles in grade II and III gliomas, termed LGGs in this study, remain to be fully elucidated. Our research aims to develop immune-related features for risk stratification and prognosis prediction in LGG. Methods: Using the ssGSEA method, we assessed the immune characteristics of the LGG population. We conducted differential analysis using LGG samples from the TCGA database and normal samples from GTEx, identifying 412 differentially expressed immune-related genes (DEIRGs). Subsequently, we utilized univariate Cox, LASSO, and multivariate Cox regression analyses to establish both a gene predictive model and a nomogram predictive model. Results: Here, we found that the ESTIMATE score, immune score and stromal score of high-immunity, high-grade and isocitrate dehydrogenase (IDH) wild-type glioma were higher than those of the corresponding group, and the tumor purity was lower. Higher ESTIMATE scores, stromal scores and immune scores indicated a poor prognosis in patients with LGG. Our four-gene prognostic model demonstrated superior accuracy compared to other molecular features. Validation using the CGGA as a testing set and the combined TCGA and CGGA cohort confirmed its robust prognostic value. Additionally, a nomogram integrating the prognostic model and clinical variables showed enhanced predictive capability. Discussion: Our study highlights the prognostic significance of the identified four DEIRGs (KLRC3, MR1, PDIA2, and RFXAP) in LGG patients. The predictive model and nomogram developed herein offer valuable tools for personalized treatment strategies in LGG. Future research should focus on further validating these findings and exploring the functional roles of these DEIRGs within the LGG tumor microenvironment. [ABSTRACT FROM AUTHOR]

Background: In sepsis, the relationship between lymphocyte counts and patient outcomes is complex. Lymphocytopenia and lymphocytosis significantly influence survival, illustrating the dual functionality of lymphocytes in responding to infections. This study investigates this complex interaction, focusing on how variations in lymphocyte counts correlate with all-cause hospital mortality among sepsis patients. Methods: This retrospective cohort study analyzed data from two extensive critical care databases: the Medical Information Mart for Intensive Care IV 2.0 (MIMIC-IV 2.0) from Beth Israel Deaconess Medical Center, Boston, Massachusetts, and the eICU Collaborative Research Database (eICU-CRD), which was Multi-center database from over 200 hospitals across the United States conducted by Philips eICU Research Institute. We included adult patients aged 18 years and older who met the Sepsis-3 criteria, characterized by documented or suspected infection and a Sequential Organ Failure Assessment (SOFA) score of 2 or higher. Sepsis patients were categorized into quartiles based on lymphocyte counts. The primary outcome was all-cause mortality in the hospital, with 90 and 60-day all-cause mortality as the secondary outcomes. Univariable and multivariable Cox proportional hazard regressions were utilized to assess lymphocyte counts' impact on hospital mortality. An adjusted restricted cubic spline (RCS) analysis was performed to elucidate this relationship further. Subgroup analyses were also conducted to explore the association across various comorbidity groups among sepsis and septic shock patients. Results: Our study included 37,054 patients, with an observed in-hospital mortality rate of 16.6%. Univariable and multivariable Cox proportional hazard regression models showed that lymphocyte counts were independently associated with in-hospital mortality (HR = 1.04, P < 0.01; HR = 1.06, P < 0.01). RCS regression analysis revealed a U-shaped relationship between lymphocyte levels and hospital mortality risk in sepsis and septic shock patients (P for overall < 0.001, P for nonliner < 0.01; P for overall = 0.002, P for nonliner = 0.014). Subgroup analyses revealed that elevated lymphocyte counts correlated with increased hospital mortality among sepsis patients with liver disease and requiring renal replacement therapy (P for overall = 0.021, P for nonliner = 0.158; P for overall = 0.025, P for nonliner = 0.759). These findings suggest that lymphocytes may have enhanced prognostic value in specific subsets of critically ill sepsis patients. Conclusion: Our findings demonstrate that lymphocyte counts are a significant independent predictor of hospital mortality in sepsis and septic shock patients. We observed a U-shaped association between lymphocyte levels and mortality risk, indicating that high and low counts are linked to increased mortality. This result highlights the complex role of lymphocytes in sepsis outcomes and suggests the need for further investigation into the underlying mechanisms and potential therapeutic approaches. Integrating lymphocyte count assessment into risk stratification algorithms and clinical decision support tools could enhance the early identification of high-risk sepsis patients. [ABSTRACT FROM AUTHOR]

The rheological parameters of oil testing fluids play a crucial role in safely conducting oil testing operations, protecting reservoirs, and maintaining wellbore integrity, which are greatly influenced by temperature and pressure. This article conducted rheological experiments on commonly used potassium salt polymer systems and organic salt systems for oil testing fluids in the Bohai Bay Oilfield under 4–200 °C and 30–150 MPa. The experiments revealed the variation of rheological parameters with temperature and pressure. Nine commonly used rheological models were evaluated for their applicability to the two types of oil testing fluids. It was found that the Ross model was suitable for the potassium salt polymer system, while the power-law model was suitable for organic salt oil testing fluids. A new rheological parameter prediction model suitable for wide range of temperature and pressure and different testing fluid systems was proposed based on the optimized results. The average error for the potassium salt polymer system for oil testing fluids was 0.101%, and the error for the organic salt system for oil testing fluids was 0.113%. Both were superior to known rheological models and can meet the requirements of oil testing operations within the wide temperature and pressure range in the Bohai Bay Oilfield. [ABSTRACT FROM AUTHOR]

The spatiotemporal mobility patterns and next location prediction of fake base stations (FBS) provide important technical support for the police to prevent spam messages from FBS. However, due to the difficulty in locating their real-time locations, our understanding of the mobility patterns and predictability of FBS is still limited. Based on the crowdsourced spam data, we extract the time and potential locations of FBS and propose a Tucker-MMC method that combines Tucker decomposition with a Mobility Markov Chain (MMC) model to investigate the mobility patterns and predictability of FBS sending spam messages. First, we utilize Tucker decomposition to reflect the spatial and temporal preferences during the movement of the corresponding FBS. Then the mobility regularity and the theoretical maximum predictability of the FBS trajectories with similar mobility preferences are analyzed by entropy and Fano's inequality. A Tucker-MMC is also established for the next location prediction. The results using the spam dataset in Beijing show that the accuracy of Tucker-MMC is more than double that of the MMC. The accuracy of the actual location prediction model is more likely to approach the theoretical maximum predictability when FBS send spam messages in a shorter time, shorter transfer distance, and smaller access range. [ABSTRACT FROM AUTHOR]

Congenital long QT syndrome (LQTS) diagnosis is complicated by limited genetic testing at scale, low prevalence, and normal QT corrected interval in patients with high-risk genotypes. We developed a deep learning approach combining electrocardiogram (ECG) waveform and electronic health record data to assess whether patients had pathogenic variants causing LQTS. We defined patients with high-risk genotypes as having ≥1 pathogenic variant in one of the LQTS-susceptibility genes. We trained the model using data from United Kingdom Biobank (UKBB) and then fine-tuned in a racially/ethnically diverse cohort using Mount Sinai BioMe Biobank. Following group-stratified 5-fold splitting, the fine-tuned model achieved area under the precision-recall curve of 0.29 (95% confidence interval [CI] 0.28–0.29) and area under the receiver operating curve of 0.83 (0.82–0.83) on independent testing data from BioMe. Multimodal fusion learning has promise to identify individuals with pathogenic genetic mutations to enable patient prioritization for further work up. [ABSTRACT FROM AUTHOR]

Clostridioides difficile infection (CDI) is a major cause of antibiotic-associated diarrhea and colitis. It is recognized as one of the most significant hospital-acquired infections. Although CDI can develop severe complications and spores of Clostridioides difficile can be transmitted by the fecal-oral route, CDI is occasionally overlooked in clinical settings. Thus, it is necessary to monitor high CDI risk groups, particularly those undergoing antibiotic treatment, to prevent complications and spread. We developed and validated a deep learning-based model to predict the occurrence of CDI within 28 days after starting antibiotic treatment using longitudinal electronic health records. For each patient, timelines of vital signs and laboratory tests with a 35-day monitoring period and a patient information vector consisting of age, sex, comorbidities, and medications were constructed. Our model achieved the prediction performance with an area under the receiver operating characteristic curve of 0.952 (95% CI: 0.932–0.973) in internal validation and 0.972 (95% CI: 0.968–0.975) in external validation. Platelet count and body temperature emerged as the most important features. The risk score, the output value of the model, exhibited a consistent increase in the CDI group, while the risk score in the non-CDI group either maintained its initial value or decreased. Using our CDI prediction model, high-risk patients requiring symptom monitoring can be identified. This could help reduce the underdiagnosis of CDI, thereby decreasing transmission and preventing complications. [ABSTRACT FROM AUTHOR]

Background: The rise of network pharmacology has led to the widespread use of network-based computational methods in predicting drug target interaction (DTI). However, existing DTI prediction models typically rely on a limited amount of data to extract drug and target features, potentially affecting the comprehensiveness and robustness of features. In addition, although multiple networks are used for DTI prediction, the integration of heterogeneous information often involves simplistic aggregation and attention mechanisms, which may impose certain limitations. Results: MSH-DTI, a deep learning model for predicting drug-target interactions, is proposed in this paper. The model uses self-supervised learning methods to obtain drug and target structure features. A Heterogeneous Interaction-enhanced Feature Fusion Module is designed for multi-graph construction, and the graph convolutional networks are used to extract node features. With the help of an attention mechanism, the model focuses on the important parts of different features for prediction. Experimental results show that the AUROC and AUPR of MSH-DTI are 0.9620 and 0.9605 respectively, outperforming other models on the DTINet dataset. Conclusion: The proposed MSH-DTI is a helpful tool to discover drug-target interactions, which is also validated through case studies in predicting new DTIs. [ABSTRACT FROM AUTHOR]

Background: This study aimed to develop prognostic models for predicting the need for invasive mechanical ventilation (IMV) in intensive care unit (ICU) patients with COVID-19 and compare their performance with the Respiratory rate-OXygenation (ROX) index. Methods: A retrospective cohort study was conducted using data collected between March 2020 and August 2021 at three hospitals in Rio de Janeiro, Brazil. ICU patients aged 18 years and older with a diagnosis of COVID-19 were screened. The exclusion criteria were patients who received IMV within the first 24 h of ICU admission, pregnancy, clinical decision for minimal end-of-life care and missing primary outcome data. Clinical and laboratory variables were collected. Multiple logistic regression analysis was performed to select predictor variables. Models were based on the lowest Akaike Information Criteria (AIC) and lowest AIC with significant p values. Assessment of predictive performance was done for discrimination and calibration. Areas under the curves (AUC)s were compared using DeLong's algorithm. Models were validated externally using an international database. Results: Of 656 patients screened, 346 patients were included; 155 required IMV (44.8%), 191 did not (55.2%), and 207 patients were male (59.8%). According to the lowest AIC, arterial hypertension, diabetes mellitus, obesity, Sequential Organ Failure Assessment (SOFA) score, heart rate, respiratory rate, peripheral oxygen saturation (SpO2), temperature, respiratory effort signals, and leukocytes were identified as predictors of IMV at hospital admission. According to AIC with significant p values, SOFA score, SpO2, and respiratory effort signals were the best predictors of IMV; odds ratios (95% confidence interval): 1.46 (1.07–2.05), 0.81 (0.72–0.90), 9.13 (3.29–28.67), respectively. The ROX index at admission was lower in the IMV group than in the non-IMV group (7.3 [5.2–9.8] versus 9.6 [6.8–12.9], p < 0.001, respectively). In the external validation population, the area under the curve (AUC) of the ROX index was 0.683 (accuracy 63%), the AIC model showed an AUC of 0.703 (accuracy 69%), and the lowest AIC model with significant p values had an AUC of 0.725 (accuracy 79%). Conclusions: In the development population of ICU patients with COVID-19, SOFA score, SpO2, and respiratory effort signals predicted the need for IMV better than the ROX index. In the external validation population, although the AUCs did not differ significantly, the accuracy was higher when using SOFA score, SpO2, and respiratory effort signals compared to the ROX index. This suggests that these variables may be more useful in predicting the need for IMV in ICU patients with COVID-19. ClinicalTrials.gov identifier: NCT05663528. [ABSTRACT FROM AUTHOR]

Background: Cardiovascular diseases (CVD) affect over half a billion people worldwide and are the leading cause of global deaths. In particular, due to population aging and worldwide spreading of risk factors, the prevalence of heart failure (HF) is also increasing. HF accounts for approximately 36% of all CVD-related deaths and stands as the foremost cause of hospitalization. Patients affected by CVD or HF experience a substantial decrease in health-related quality of life compared to healthy subjects or affected by other diffused chronic diseases. Main body: For both CVD and HF, prediction models have been developed, which utilize patient data, routine laboratory and further diagnostic tests. While some of these scores are currently used in clinical practice, there still is a need for innovative approaches to optimize CVD and HF prediction and to reduce the impact of these conditions on the global population. Epigenetic biomarkers, particularly DNA methylation (DNAm) changes, offer valuable insight for predicting risk, disease diagnosis and prognosis, and for monitoring treatment. The present work reviews current information relating DNAm, CVD and HF and discusses the use of DNAm in improving clinical risk prediction of CVD and HF as well as that of DNAm age as a proxy for cardiac aging. Conclusion: DNAm biomarkers offer a valuable contribution to improving the accuracy of CV risk models. Many CpG sites have been adopted to develop specific prediction scores for CVD and HF with similar or enhanced performance on the top of existing risk measures. In the near future, integrating data from DNA methylome and other sources and advancements in new machine learning algorithms will help develop more precise and personalized risk prediction methods for CVD and HF. [ABSTRACT FROM AUTHOR]

Background: This study aims to analyse changes in urinary kidney injury markers in children with Mycoplasma pneumoniae pneumonia (MPP), investigate the risk factors for MPP-related acute kidney injury (AKI) and establish a model to predict MPP-related AKI. Methods: Ninety-five children were enrolled based on the study's inclusion and exclusion criteria. They were divided into a severe MPP (SMPP) group and a non-SMPP group and then into an AKI group and a non-AKI group according to the presence of AKI. A univariate logistic regression analysis was performed to explore the early risk factors for AKI. Based on a multivariate logistic regression analysis and a least absolute shrinkage and selection operator regression analysis, appropriate variables were selected to establish a prediction model, and R 4.2.2 software was used to draw nomograms and generate a dynamic nomogram website. Results: Seven urinary kidney injury markers were abnormally elevated in the SMPP group and the non-SMPP group: urinary N-acetyl-β-D-glucosaminidase (NAG), β2-microglobulin, α1-microglobulin, retinol-binding protein, urinary immunoglobulin G, urinary transferrin and urinary microalbumin. Sixteen children were identified with AKI during hospitalisation. The AKI group had higher levels of urinary NAG, α1-microglobulin, β2-microglobulin, urinary microalbumin, urinary transferrin and retinol-binding protein than the non-AKI group (P < 0.05). The MPP-related AKI prediction model consists of four indicators (serum immunoglobulin M [IgM], C-reactive protein [CRP], urine NAG and sputum plug presence) and a dynamic nomogram. Conclusion: Urinary kidney injury markers are often elevated in children with MPP; urinary NAG is the marker most likely to be elevated, and it is especially evident in severe cases. The nomogram of the prediction model, comprising serum IgM, CRP, urinary NAG and sputum plug presence, can predict the probability of AKI in children with MPP. [ABSTRACT FROM AUTHOR]

Background: The Achilles tendon is the body's strongest and largest tendon. It is commonly injured, particularly among athletes, accounting for a significant portion of serious tendon injuries. Several factors play a precipitating role in increasing the risk of these injuries. Objective: Our objective is to derive and validate a risk calculator for the prediction of incidence of any complication following Achilles tendon repair. Methods: We used de-identified data from the American College of Surgeons' National Surgical Quality Improvement Project (NSQIP) database from 2005 to 2021. It comprises 7010 individuals who had undergone Achilles tendon rupture repair. Demographic and risk factors information was collected. To develop the calculator, the sample was divided into a derivation cohort (40%) and a validation cohort (60%). Multivariate logistic regression was used for statistical analysis, and a risk calculator for incidence of any complication was derived from the derivation cohort and validated on the remaining 60% of the sample. Patients with missing data were excluded, and the significance level was set at p < 0.05. Results: We analyzed the derivation cohort of 2245 individuals who underwent Achilles tendon repair surgery between 2005 and 2021, with a 5.5% overall complication. Multivariate logistic regression identified anesthesia type, ASA classification, certain co-morbidities (pre-operative dialysis and medication-requiring hypertension), and wound classification as significant predictors of complications. The developed risk calculator model had an area under the curve (AUC) of 0.685 in the derivation cohort and 0.655 in the validation cohort, surpassing the widely used and validated modified frailty index. A cut-off score threshold of 0.06 was established using Youden's index to dichotomize individuals into low and high risk for developing any postoperative complications. Conclusion: Our risk calculator includes factors that most significantly affect the incidence of any complication following Achilles tendon repair. [ABSTRACT FROM AUTHOR]

Objectives: We aimed to evaluate the ability of the parasternal intercostal (PIC) thickening fraction during spontaneous breathing trial (SBT) to predict the need for reintubation within 48 h after extubation in surgical patients with sepsis. Methods: This prospective observational study included adult patients with sepsis who were mechanically ventilated and indicated for SBT. Ultrasound measurements of the PIC thickening fraction and diaphragmatic excursion (DE) were recorded 15 min after the start of the SBT. After extubation, the patients were followed up for 48 h for the need for reintubation. The study outcomes were the ability of the PIC thickening fraction (primary outcome) and DE to predict reintubation within 48 h of extubation using area under receiver characteristic curve (AUC) analysis. The accuracy of the model including the findings of right PIC thickening fraction and right DE was also assessed using the current study cut-off values. Multivariate analysis was performed to identify independent risk factors for reintubation. Results: We analyzed data from 49 patients who underwent successful SBT, and 10/49 (20%) required reintubation. The AUCs (95% confidence interval [CI]) for the ability of right and left side PIC thickening fraction to predict reintubation were 0.97 (0.88–1.00) and 0.96 (0.86–1.00), respectively; at a cutoff value of 6.5–8.3%, the PIC thickening fraction had a negative predictive value of 100%. The AUCs for the PIC thickening fraction and DE were comparable; and both measures were independent risk factors for reintubation. The AUC (95% CI) of the model including the right PIC thickening fraction > 6.5% and right DE ≤ 18 mm to predict reintubation was 0.99 (0.92–1.00), with a positive predictive value of 100% when both sonographic findings are positive and negative predictive value of 100% when both sonographic findings are negative. Conclusions: Among surgical patients with sepsis, PIC thickening fraction evaluated during the SBT is an independent risk factor for reintubation. The PIC thickening fraction has an excellent predictive value for reintubation. A PIC thickening fraction of ≤ 6.5–8.3% can exclude reintubation, with a negative predictive value of 100%. Furthermore, a combination of high PIC and low DE can also indicate a high risk of reintubation. However, larger studies that include different populations are required to replicate our findings and validate the cutoff values. [ABSTRACT FROM AUTHOR]

Introduction Material and Methods Results Conclusions This study aimed to validate the Sargent risk stratification algorithm for the prediction of placenta accreta spectrum (PAS) severity using data collected from multiple centers and using the multicenter data to improve the model.We conducted a multicenter analysis using data collected for the IS‐PAS database. The Sargent model's effectiveness in distinguishing between abnormally adherent placenta (FIGO grade 1) and abnormally invasive placenta (FIGO grades 2 and 3) was evaluated. A new model was developed using multicenter data from the IS‐PAS database.The database included 315 cases of suspected PAS, of which 226 had fully documented standardized ultrasound signs. The final diagnosis was normal placentation in 5, abnormally adherent placenta/FIGO grade 1 in 43, and abnormally invasive placenta/FIGO grades 2 and 3 in 178. The external validation of the Sargent model revealed moderate predictive accuracy in a multicenter setting (C‐index 0.68), compared to its higher accuracy in a single‐center context (C‐index 0.90). The newly developed model achieved a C‐index of 0.74.The study underscores the difficulty in developing universally applicable PAS prediction models. While models like that of Sargent et al. show promise, their reproducibility varies across settings, likely due to the interpretation of the ultrasound signs. The findings support the need for updating the current ultrasound descriptors and for the development of any new predictive models to use data collected by different operators in multiple clinical settings. [ABSTRACT FROM AUTHOR]

Load prediction tasks aim to predict the dynamic trend of future load based on historical performance sequences, which are crucial for cloud platforms to make timely and reasonable task scheduling. However, existing prediction models are limited while capturing complicated temporal patterns from the load sequences. Besides, the frequently adopted global weighting strategy (e.g., the self-attention mechanism) in temporal modeling schemes has quadratic computational complexity, hindering the immediate response of cloud servers in complex real-time scenarios. To address the above limitations, we propose a Wavelet decomposition-enhanced External Transformer (WETformer) to provide accurate yet efficient load prediction for cloud servers. Specifically, we first incorporate discrete wavelet transform to progressively extract long-term trends, highlighting the intrinsic attributes of temporal sequences. Then, we propose a lightweight multi-head External Attention (EA) mechanism to simultaneously consider the inter-element relationships within load sequences and the correlations across different sequences. Such an external component has linear computational complexity, mitigating the encoding redundancy prevalent and enhancing prediction efficiency. Extensive experiments conducted on Alibaba Cloud's cluster tracking dataset demonstrate that WETformer achieves superior prediction accuracy and the shortest inference time compared to several state-of-the-art baseline methods. [ABSTRACT FROM AUTHOR]

Perceptual decisions should depend on sensory evidence. However, such decisions are also influenced by past choices and outcomes. These choice history biases may reflect advantageous strategies to exploit temporal regularities of natural environments. However, it is unclear whether and how observers can adapt their choice history biases to different temporal regularities, to exploit the multitude of temporal correlations that exist in nature. Here, we show that male mice adapt their perceptual choice history biases to different temporal regularities of visual stimuli. This adaptation was slow, evolving over hundreds of trials across several days. It occurred alongside a fast non-adaptive choice history bias, limited to a few trials. Both fast and slow trial history effects are well captured by a normative reinforcement learning algorithm with multi-trial belief states, comprising both current trial sensory and previous trial memory states. We demonstrate that dorsal striatal dopamine tracks predictions of the model and behavior, suggesting that striatal dopamine reports reward predictions associated with adaptive choice history biases. Our results reveal the adaptive nature of perceptual choice history biases and shed light on their underlying computational principles and neural correlates. The world exhibits temporal regularities that can be exploited to improve perceptual decisions. Here, the authors show that mice adapt to such regularities, well described by reinforcement learning, and that dopamine release in the striatum tracks this adaptation. [ABSTRACT FROM AUTHOR]

In this article, a model predictive control (MPC) with common-mode voltage (CMV) suppression is proposed for single-phase cascaded H-bridge (CHB) inverters, which can also simultaneously achieve control objectives of gridconnected current tracking, voltages balancing of different H-bridge submodules on the DC-side and switching frequency reduction. To suppress high-frequency components of the common-mode voltage without additional switching devices, the algorithm proposed designs the predicted and reference values of the CMV and incorporates them in the cost function. At the same time, the capacitor voltages balancing control is integrated in the calculation of the optimal modulation function of the H-bridge, which reduces the complexity of control effectively. Besides, switching times of the MOSFETs are compared in two cycles. The cost function is constructed to represent comprehensive effect of the control. Finally, an experiment is performed on the hardware-in-the-loop experimental platform. The experimental results show that the proposed algorithm can offer a better voltage THD and reduce the times of switch action by nearly half while maintaining high-precision current tracking and maximum power point of photovoltaic modules, which alleviate the potential electromagnetic interference and cabling problem. [ABSTRACT FROM AUTHOR]

This study proposes a cooperative lane‐changing control framework for multiple vehicles in freeway ramp merging areas, aiming to achieve safe and efficient merging. Specifically, multiple connected and automated vehicles (CAVs) form triplets to participate in cooperative lane‐changing. The framework consists of two stages: Longitudinal Headway Adjustment (LHA) and Lane‐Changing Execution (LCE). In the LHA stage, a centralized longitudinal controller is developed based on the vehicle's longitudinal dynamics model to optimize the longitudinal velocity of the cooperative vehicles and create suitable gaps for merging vehicles. In the LCE stage, an optimal lane‐changing reference trajectory is generated using a quintic polynomial and a lateral controller is designed based on the vehicle's lateral dynamics model. Model Predictive Control (MPC) is utilized for trajectory tracking. The simulation results obtained using MATLAB/Simulink, GPOPS, and CarSim demonstrate that the proposed control strategy can be applied to different vehicle speed control scenarios. Taking a specific velocity combination as an example, the cumulative control errors in the longitudinal and lateral directions for PV (Preceding Vehicle), SV (Subject Vehicle), and FV (Following Vehicle) are 1.4014 m, 0.5631 m, and −0.7601 m, respectively, satisfying the safety distance requirements. Compared to the Linear Quadratic Regulator (LQR) control, the proposed strategy improves control efficiency by 145.03%, 69.64%, 43.18%, and 67.61% in terms of comprehensive spacing errors, synthesized acceleration, front wheel angle, and speed fluctuation, respectively. These research findings highlight the superior performance of the proposed control strategy in terms of traffic efficiency, comfort, safety, and vehicle stability. [ABSTRACT FROM AUTHOR]

Purpose: To identify the key risk factors for venous thromboembolism (VTE) in urological inpatients based on the Caprini scale using an interpretable machine learning method. Methods: VTE risk data of urological inpatients were obtained based on the Caprini scale in the case hospital. Based on the data, the Boruta method was used to further select the key variables from the 37 variables in the Caprini scale. Furthermore, decision rules corresponding to each risk level were generated using the rough set (RS) method. Finally, random forest (RF), support vector machine (SVM), and backpropagation artificial neural network (BPANN) were used to verify the data accuracy and were compared with the RS method. Results: Following the screening, the key risk factors for VTE in urology were "(C1) Age," "(C2) Minor Surgery planned," "(C3) Obesity (BMI > 25)," "(C8) Varicose veins," "(C9) Sepsis (< 1 month)," (C10) "Serious lung disease incl. pneumonia (< 1month) " (C11) COPD," "(C16) Other risk," "(C18) Major surgery (> 45 min)," "(C19) Laparoscopic surgery (> 45 min)," "(C20) Patient confined to bed (> 72 h)," "(C18) Malignancy (present or previous)," "(C23) Central venous access," "(C31) History of DVT/PE," "(C32) Other congenital or acquired thrombophilia," and "(C34) Stroke (< 1 month." According to the decision rules of different risk levels obtained using the RS method, "(C1) Age," "(C18) Major surgery (> 45 minutes)," and "(C21) Malignancy (present or previous)" were the main factors influencing mid- and high-risk levels, and some suggestions on VTE prevention were indicated based on these three factors. The average accuracies of the RS, RF, SVM, and BPANN models were 79.5%, 87.9%, 92.6%, and 97.2%, respectively. In addition, BPANN had the highest accuracy, recall, F1-score, and precision. Conclusions: The RS model achieved poorer accuracy than the other three common machine learning models. However, the RS model provides strong interpretability and allows for the identification of high-risk factors and decision rules influencing high-risk assessments of VTE in urology. This transparency is very important for clinicians in the risk assessment process. [ABSTRACT FROM AUTHOR]

Debris flows and floods generated by rainfall runoff occur in rocky mountainous landscapes and burned steeplands. Flow type is commonly identified post‐event through interpretation of depositional structures, but these may be poorly preserved or misinterpreted. Prior research indicates that discharge magnitude is commonly amplified in debris flows relative to floods due to volumetric bulking and increased frictional resistance. Here, we use this flow amplification to develop a metric (Q*) to separate debris flows from floods based on the ratio of observed peak discharge to the theoretical maximum water discharge from rainfall runoff. We compile 642 observations of floods and debris flows and demonstrate that Q* distinguishes flow type to ∼92% accuracy. Q* allows for accurate identification of debris flows through simple channel cross‐section surveys rather than through qualitative interpretation of deposits, and therefore should increase the performance of models and engineered structures that require accurate flow‐type observations. Plain Language Summary: Steep channels in rocky landscapes and recently burned areas carry a wide range of flow types, ranging from water floods to highly destructive debris flows composed of rock, sediment, other incidental debris, and water. Typically, the flow type that recently occurred in a channel is identified by interpreting the structure of deposits left behind by the flow. However, these deposits can be challenging to interpret, provided they are preserved at all. Prior research has shown that the shift in flow physics occurring in debris flows has the effect of amplifying the magnitude of the flow beyond that which is possible in floods. In this study, we take advantage of this flow amplification to develop a quantitative metric that allows for flow type to be identified without interpreting depositional structures. We demonstrate the metric to be highly accurate. This methodology should increase the accuracy of predictive models trained on such observations, which can help to better understand where these hazards exist and protect communities and infrastructure at risk. Key Points: To distinguish debris flows from floods, a metric Q* equal to the ratio of peak discharge to rainfall‐runoff discharge is developedQ* separates floods (Q* ≤ 1) from debris flows (Q* > 1) to ∼92% accuracy and is robust to uncertainties in the key parametersThe Q* metric should reduce misidentification of flow type and increase the performance of empirical‐hazard models and hazard mitigation [ABSTRACT FROM AUTHOR]

Background: Regions of genome-wide marker data may have differing influences on the evaluated traits. This can be reflected in the genomic models by assigning different weights to the markers, which can enhance the accuracy of genomic prediction. However, the standard multi-trait single-step genomic evaluation model can be computationally infeasible when the traits are allowed to have different marker weights. Results: In this study, we developed and implemented a multi-trait single-step single nucleotide polymorphism best linear unbiased prediction (SNPBLUP) model for large genomic data evaluations that allows for the use of precomputed trait-specific marker weights. The modifications to the standard single-step SNPBLUP model were minor and did not significantly increase the preprocessing workload. The model was tested using simulated data and marker weights precomputed using BayesA. Based on the results, memory requirements and computing time per iteration slightly increased compared to the standard single-step model without weights. Moreover, convergence of the model was slower when using marker weights, which resulted in longer total computing time. The use of marker weights, however, improved prediction accuracy. Conclusions: We investigated a single-step SNPBLUP model that can be used to accommodate trait-specific marker weights. The marker-weighted single-step model improved prediction accuracy. The approach can be used for large genomic data evaluations using precomputed marker weights. [ABSTRACT FROM AUTHOR]

Deep learning has enabled breakthroughs in automated diagnosis from medical imaging, with many successful applications in ophthalmology. However, standard medical image classification approaches only assess disease presence at the time of acquisition, neglecting the common clinical setting of longitudinal imaging. For slow, progressive eye diseases like age-related macular degeneration (AMD) and primary open-angle glaucoma (POAG), patients undergo repeated imaging over time to track disease progression and forecasting the future risk of developing a disease is critical to properly plan treatment. Our proposed Longitudinal Transformer for Survival Analysis (LTSA) enables dynamic disease prognosis from longitudinal medical imaging, modeling the time to disease from sequences of fundus photography images captured over long, irregular time periods. Using longitudinal imaging data from the Age-Related Eye Disease Study (AREDS) and Ocular Hypertension Treatment Study (OHTS), LTSA significantly outperformed a single-image baseline in 19/20 head-to-head comparisons on late AMD prognosis and 18/20 comparisons on POAG prognosis. A temporal attention analysis also suggested that, while the most recent image is typically the most influential, prior imaging still provides additional prognostic value. [ABSTRACT FROM AUTHOR]

Tunnel water inrush may not only cause hundreds of millions of economic losses and serious casualties, but also leads to a series of ecological and environmental problems such as the decline of groundwater level, soil salinization and surface vegetation degradation. In this study, considering hydrogeology, construction, and dynamic monitoring factors, a new risk prediction model of water inrush is proposed based on fuzzy mathematical theory. The element of novelty is that this approach comprehensively considers nonlinearity and randomness factors, and the index values, weights, and membership are expressed as interval numbers instead of constant values. The interval membership degree of each index is calculated by an improved sigmoid membership function (SMF). A coupling algorithm of improved analytic hierarchy process and entropy method is used to calculate the index weight. In addition, the Boolean matrix is introduced into the relative advantage analysis of the interval vector, and the final risk level of water inrush is determined by the ranking result. The proposed model is applied to the analysis of the water inrush risk in the Ka-Shuang 2 (KS2) tunnel in Xinjiang, China. The predicted results align well with the actual excavation results, which indicates that this novel model has high accuracy and reliability. Simultaneously, a risk management response mechanism for different risk levels of water inrush is discussed, which is expected to provide a new research perspective for risk control of other related projects and promote regional sustainable development. [ABSTRACT FROM AUTHOR]

Background: Accurate prediction of postoperative vault in implantable collamer lens (ICL) implantation is crucial; however, current formulas often fail to account for individual anatomical variations, leading to suboptimal visual outcomes and necessitating improved predictive models. We aimed to verify the prediction accuracy of our new predictive model for vaulting based on anterior and posterior chamber structural parameters. Methods: This retrospective observational study included 137 patients (240 eyes) who previously underwent ICL surgery. Patients were randomly divided into the model establishment (192 eyes) or validation (48 eyes) groups. Preoperative measurements of the anterior and posterior chamber structures were obtained using Pentacam, CASIA2 anterior segment optical coherence tomography (AS-OCT), ultrasound biomicroscopy, and other devices. Stepwise multiple linear regression analysis was used to evaluate the relationship between the vault and each variable (WL formula). The Friedman test was performed for the vaulting prediction results of the WL, NK (Ver. 3), and KS formulas (Ver. 4) in CASIA2 AS-OCT, as well as the Zhu formula and vault measurements. The proportions of prediction error within ± 250 μm per formula were compared. Results: The predicted vault values of the WL, NK, KS, and Zhu formulas and vault measurements were 668.74 ± 162.12, 650.85 ± 248.47, 546.56 ± 128.99, 486.56 ± 210.76, and 716.06 ± 233.84 μm, respectively, with a significant difference (χ2 = 69.883, P = 0.000). Significant differences were also found between the measured vault value and Zhu formula, measured vault value and KS formula, WL formula and Zhu formula, WL formula and KS formula, NK formula and KS formula, and NK formula and Zhu formula (P < 0.001) but not between other groups. The proportions of prediction error within ± 250 μm per formula were as follows: WL formula (81.3%) > NK formula (70.8%) > KS formula (66.7%) > Zhu formula (54.2%). Conclusions: The WL formula, which considers the complexity of the anterior and posterior chamber structures, demonstrates greater calculation accuracy, compared with the KS (Ver. 4) and Zhu formulas. The proportion of absolute prediction error ≤ 250 μm is higher with the WL formula than with the NK formula (ver. 3). This enhanced predictive capability can improve ICL sizing decisions, thereby increasing the safety and efficacy of ICL implantation surgeries. [ABSTRACT FROM AUTHOR]

This paper formulates a cooperative traffic control methodology that integrates traffic signal timing and ramp metering decisions into an optimization model to improve traffic operations in a corridor network. A mixed integer linear model is formulated and is solved in real time within a model predictive controller framework, where the cell transmission model is used as the system state predictor. The methodology is benchmarked in a case study corridor in San Mateo, CA, in VISSIM with two optimization scenarios, namely, optimal metering and optimal signal control, and two simulation scenarios with a preset metering plan and no metering. The numerical results show that integrated traffic signal and ramp metering control reduces delays, stops, and travel times of the corridor by up to 33.1%, 36%, and 16.4%, respectively, compared to existing benchmark conditions. With appropriate weights prioritizing freeway or arterial street operations, the integrated control balances traffic congestion between the arterial street and the freeway. [ABSTRACT FROM AUTHOR]

This paper proposes a data-driven prediction scheme for the remaining life of centrifugal pump bearings based on the KPCA–LSTM network. A centrifugal pump bearing fault experiment bench is built to collect data, and the performance of time domain, frequency domain, and time-frequency domain characteristics under different working conditions is analyzed. Time domain characteristics, frequency domain characteristics, wavelet packet decomposition energy characteristics, and CEEMDAN energy features are found to be able to capture fault information under different working conditions. Therefore, 43 sensitive features are determined from the time domain, frequency domain, and time-frequency domain. Through the analysis of XJTU-SY bearing life cycle data and based on the weighted scores of monotonicity, robustness, and trend indicators, twelve outstanding characteristics of the bearing in the whole life cycle are selected, and a one-dimensional feature quantity that can characterize the life-degradation process of the centrifugal pump bearing is constructed after KPCA dimension reduction processing. The LSTM network, sensitive to time series, is selected to predict and analyze the constructed one-dimensional feature trend, and the prediction effects of the BP network and the CNN network are compared. The results show that this method has advantages in prediction accuracy and model training time. [ABSTRACT FROM AUTHOR]

The intermittent and stochastic nature of Renewable Energy Sources (RESs) necessitates accurate power production prediction for effective scheduling and grid management. This paper presents a comprehensive review conducted with reference to a pioneering, comprehensive, and data-driven framework proposed for solar Photovoltaic (PV) power generation prediction. The systematic and integrating framework comprises three main phases carried out by seven main comprehensive modules for addressing numerous practical difficulties of the prediction task: phase I handles the aspects related to data acquisition (module 1) and manipulation (module 2) in preparation for the development of the prediction scheme; phase II tackles the aspects associated with the development of the prediction model (module 3) and the assessment of its accuracy (module 4), including the quantification of the uncertainty (module 5); and phase III evolves towards enhancing the prediction accuracy by incorporating aspects of context change detection (module 6) and incremental learning when new data become available (module 7). This framework adeptly addresses all facets of solar PV power production prediction, bridging existing gaps and offering a comprehensive solution to inherent challenges. By seamlessly integrating these elements, our approach stands as a robust and versatile tool for enhancing the precision of solar PV power prediction in real-world applications. [ABSTRACT FROM AUTHOR]

To improve the reliability and maintainability of the nuclear safety-class digital control system (DCS), this paper conducts a study on the fault prediction of critical components in the output circuit of the nuclear safety-class signal conditioning module. To address the issue of insufficient feature extraction for the minor offset fault feature and the low accuracy of fault prediction, a predictive model based on stacked denoising autoencoder (SDAE) feature extraction and an improved gated recurrent unit (GRU) is proposed. Therefore, fault simulation modeling is performed for critical components of the signal output circuit to obtain fault datasets of critical components, and the SDAE model is used to extract fault features. The fault prediction model based on GRU is established, and the number of hidden layers, the number of hidden layer nodes, and the learning rate of the GRU model are optimized using the adaptive gray wolf optimization algorithm (AGWO). The prediction performance evaluation metrics include the root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and absolute error (EA), which are used for evaluating the prediction results of models such as the AGWO-GRU model, recurrent neural network (RNN) model, and long short-term memory network (LSTM). The results show that the GRU model optimized by AGWO has a better prediction accuracy (errors range within 0.01%) for the faults of the circuit critical components, and, moreover, can accurately and stably predict the fault trend of the circuit. [ABSTRACT FROM AUTHOR]

The energy consumption of air-jet looms mainly comes from air compressors. Predicting the air consumption of air-jet looms for the upcoming period is significant for the variable frequency adjustment of air compressors, thereby aiding in energy saving and reducing fabric costs. This paper proposes an innovative method that combines Sliding Time Windows (STW), feature analysis, and incremental learning to improve the accuracy of short-term air consumption prediction. First, the STW method is employed during the data collection phase to enhance data reliability. Through feature analysis, significant factors affecting the air consumption of air-jet looms, beyond traditional research, are explored and incorporated into the prediction model. The experimental results indicate that the introduction of new features improved the model's R2 from 0.905 to 0.950 and reduced the MSE from 32.369 to 16.239. The STW method applied to the same random forest model increased the R2 from 0.906 to 0.950 and decreased the MSE from 32.244 to 16.239. The decision tree method, compared to the linear regression model, improved the R2 from 0.928 to 0.950 and reduced the MSE from 23.541 to 16.239, demonstrating significant predictive performance enhancement. After establishing the optimal model, incremental learning is used to continuously improve the reliability and accuracy of short-term predictions. Experiments show that the incremental learning method, compared to static models, offers better resilience and reliability when new data are collected. The proposed method significantly improves the accuracy of air consumption prediction for air-jet looms, providing strong support for the variable frequency adjustment of air compressors, and contributes to the goals of energy saving and cost reduction. The research results indicate that this method not only enhances prediction accuracy but also provides new insights and methods for future energy-saving research. [ABSTRACT FROM AUTHOR]

Photovoltaic (PV) power generation is significantly impacted by environmental factors that exhibit substantial uncertainty and volatility, posing a critical challenge for accurate PV power prediction in power system management. To address this, a parallel model is proposed for PV short-term prediction utilizing a multi-level attention mechanism. Firstly, gray relation analysis (GRA) and an improved ISODATA algorithm are used to select a dataset of similar days with comparable meteorological characteristics to the forecast day. A transformer encoder layer with multi-head attention is then used to extract long-term dependency features. Concurrently, BiGRU, optimized with a Global Attention network, is used to capture global temporal features. Feature fusion is performed using Cross Attention, calculating attention weights to emphasize significant features and enhancing feature integration. Finally, high-precision predictions are achieved through a fully connected layer. Utilizing historical PV power generation data to predict power output under various weather conditions, the proposed model demonstrates superior performance across all three climate types compared to other models, achieving more reliable predictions. [ABSTRACT FROM AUTHOR]