To the Editor — Validation and predictive performance of the PAT score

The high performance of the elbow joint angle prediction is essential on the development of the devices based on electromyography (EMG) control. The performance of the prediction depends on the feature of extraction parameters such as window length. In this paper, we evaluated the effect of the window length on the performance of the elbow-joint angle prediction. The prediction algorithm consists of zero-crossing feature extraction and second order of Butterworth low pass filter. The feature was used to extract the EMG signal by varying window length. The EMG signal was collected from the biceps muscle while the elbow was moved in the flexion and extension motion. The subject performed the elbow motion by holding a 1-kg load and moved the elbow in different periods (12 seconds, 8 seconds and 6 seconds). The results indicated that the window length affected the performance of the prediction. The 250 window lengths yielded the best performance of the prediction algorithm of (mean±SD) root mean square error = 5.68%±1.53% and Person’s correlation = 0.99±0.0059.

Classical regression is a supervised task that uses target output to guide the modeling. Generally, the original input contains both output relevant and irrelevant information, whereas the latter may decrease the predictive performance to a certain extent. Without prior information about what information contributes to the prediction, the predictive performance of regression models can be improved based on such output relevant information as input. Thus, a joint autoencoder (JAE) combining the supervised and unsupervised mechanisms is proposed to learn output relevant features from the original input. In this way, both predictive and reconstructive performance are considered to learn the essential characteristics and avoid poor generalized performance on untrained instances. Meanwhile, hierarchy representations can be learned by successive JAEs with a local parameter embedding strategy, which is presented to preserve the predictive performance in this structure. In the experiments, the predictive performance and robustness of the proposal are verified on nine datasets with different sample sizes.

Background: Heart failure poses a significant global health challenge, with high rates of readmission and mortality. Accurate models to predict these outcomes are essential for effective patient management. This study investigates the impact of data pre-processing techniques on XGBoost model performance in predicting all-cause readmission and mortality among heart failure patients. Methods: A dataset of 168 features from 2008 heart failure patients was used. Pre-processing included handling missing values, categorical encoding, and standardization. Four imputation techniques were compared: Mean, Multivariate Imputation by Chained Equations (MICEs), k-nearest Neighbors (kNNs), and Random Forest (RF). XGBoost models were evaluated using accuracy, recall, F1-score, and Area Under the Curve (AUC). Robustness was assessed through 10-fold cross-validation. Results: The XGBoost model with kNN imputation, one-hot encoding, and standardization outperformed others, with an accuracy of 0.614, recall of 0.551, and F1-score of 0.476. The MICE-based model achieved the highest AUC (0.647) and mean AUC (0.65 ± 0.04) in cross-validation. All pre-processed models outperformed the default XGBoost model (AUC: 0.60). Conclusions: Data pre-processing, especially MICE with one-hot encoding and standardization, improves XGBoost performance in heart failure prediction. However, moderate AUC scores suggest further steps are needed to enhance predictive accuracy.

Aerosol formation is a complex process involving numerous molecules, whose identities and environmental variations remain largely uncharted (Bianchi et al., 2019). Computational simulations and property prediction tools have emerged to identify compounds likely to participate in the particle formation process (Elm et al., 2020). In recent years, predictive machine learning models for saturation vapor pressure and partition coefficient estimation have achieved impressive accuracy, with mean absolute errors within one order of magnitude (Besel et al., 2023; Lumiaro et al., 2021); an advancement that enables the categorization of molecules into different volatility regions. However, the interpretability of these models in molecular sciences is often limited unless the molecular descriptor used is easily interpretable. Another challenge is that atmospheric molecules possess unique characteristics that may be overlooked by standard molecular representations developed in other chemical domains (Sandström et al., 2024). We hypothesize that combining sufficiently informative and interpretable descriptors with modern machine learning methods, chemical insight of these largely unknown chemical spaces can be gained in a data-driven way.In this contribution, we introduce a new interpretable molecular descriptor, ATMOMACCS, specifically tailored to atmospheric molecules. We demonstrate its competitive performance in predicting various thermodynamic properties, such as saturation vapor pressure, vaporization enthalpy, partition coefficients, and glass-transition temperature, equaling or surpassing published results for four distinct atmospheric molecular datasets (Besel et al., 2023; Wang et al., 2017; Ferraz-Caetano et al., 2024; Li et al., 2020). Our descriptor is based on enumerating atmospherically relevant structural motifs, making it readily interpretable for atmospheric chemists. Additionally, in our approach, we analyze the relative importance of these motifs with Shapley Additive Explanations (SHAP) values (Lundberg & Lee, 2017), providing insight into the performance improvements observed. Notably, from this analysis, we found that explicitly counting the number of carbon atoms is particularly important for property prediction, though less so for water-gas phase partition coefficients. Moreover, the analysis shows that general structural motifs are roughly equally important as motifs specific to atmospheric organic chemistry, and the combinations of these two types of motifs were pivotal for predictive performance.Our molecular descriptor, ATMOMACCS, can serve as a vital tool for advancing data-driven atmospheric science, addressing the need for more customized and accurate modelling in the field. Furthermore, the descriptor’s inherent interpretability and its strong performance in thermodynamic property prediction, with machine learning, show promise for further research in atmospheric chemistry.This work was supported by the VILMA (Virtual laboratory for molecular level atmospheric transformations) centre of excellence funded by the Academy of Finland under grant 346377. Besel, V. et al. (2023). Sci. Data 10, 1–11.Bianchi, F et al. (2019). Chem. Rev. 119, 3472–3509.Elm, J. et al. (2020) J. Aerosol Sci. 149, 105621.Ferraz-Caetano, J. et al. (2024). Chemosphere, 359, 142257Li, Y. et al. (2020). Atmos. Chem. Phys. 20, 8103–8122.Lumiaro, E. et al. (2021). Atmos. Chem. Phys. 21, 13227–13246.Lundberg, S. M. and Lee S.-I. (2017). Curran Associates Inc. 30, 9781510860964.Sandström, H. et al. (2024). Adv. Sci. 11, 2306235.Wang, C. et al. (2017).  Atmos. Chem. Phys. 17, 7529–7540.

BackgroundThis study aimed to construct predictive models for the risk of sepsis in patients with Acute pancreatitis (AP) using machine learning methods and compared optimal one with the logistic regression (LR) model and scoring systems.MethodsIn this retrospective cohort study, data were collected from the Medical Information Mart for Intensive Care III (MIMIC III) database between 2001 and 2012 and the MIMIC IV database between 2008 and 2019. Patients were randomly divided into training and test sets (8:2). The least absolute shrinkage and selection operator (LASSO) regression plus 5-fold cross-validation were used to screen and confirm the predictive factors. Based on the selected predictive factors, 6 machine learning models were constructed, including support vector machine (SVM), K-nearest neighbour (KNN), multi-layer perceptron (MLP), LR, gradient boosting decision tree (GBDT) and adaptive enhancement algorithm (AdaBoost). The models and scoring systems were evaluated and compared using sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy, and the area under the curve (AUC).ResultsA total of 1, 672 patients were eligible for participation. In the training set, 261 AP patients (19.51%) were diagnosed with sepsis. The predictive factors for the risk of sepsis in AP patients included age, insurance, vasopressors, mechanical ventilation, Glasgow Coma Scale (GCS), heart rate, respiratory rate, temperature, SpO2, platelet, red blood cell distribution width (RDW), International Normalized Ratio (INR), and blood urea nitrogen (BUN). The AUC of the GBDT model for sepsis prediction in the AP patients in the testing set was 0.985. The GBDT model showed better performance in sepsis prediction than the LR, systemic inflammatory response syndrome (SIRS) score, bedside index for severity in acute pancreatitis (BISAP) score, sequential organ failure assessment (SOFA) score, quick-SOFA (qSOFA), and simplified acute physiology score II (SAPS II).ConclusionThe present findings suggest that compared to the classical LR model and SOFA, qSOFA, SAPS II, SIRS, and BISAP scores, the machine learning model-GBDT model had a better performance in predicting sepsis in the AP patients, which is a useful tool for early identification of high-risk patients and timely clinical interventions.

Fractures of pelvis and acetabulum are at high risk of death worldwide. However, the capability of mortality prediction by the conventional system Simple Acute Physiologic Score (SAPS) II remains limited. Here, we hypothesized that the use of machine learning (ML) algorithms could provide better performance of prediction than SAPS II for traumatic patients in the intensive care unit (ICU). Customized ML models were built including support vector machine, decision tree, logistic regression and random forest based on MIMIC-III clinical database. 307 patients were enrolled with an ICD-9 diagnosis of pelvic, acetabular or combined pelvic and acetabular fractures. Feature expansion method was used for experimental models by extending SAPS II features to the first 72 hours after ICU admission, compared with the traditional first-24-hours ones that were used to build respective control models. A comparison of each model’s performance was made through the area under the receiver-operating characteristic curve (AUROC). All the ML models outperformed SAPS II system (AUROC = 0.73), among which the experimental random forest model had the supreme performance (AUROC of 0.90). The results suggested that ML models could aid in better performance of mortality prediction for pelvic and acetabular injuries and potentially support decision-making for orthopedics and ICU practitioners.

The accurate preoperative histopathological grade diagnosis of adult gliomas is of great significance for the formulation of a surgical plan and the implementation of a subsequent treatment. The aim of this study is to establish a predictive model for classifying adult gliomas into grades 2-4 based on preoperative conventional multimodal MRI radiomics. Patients with pathologically confirmed gliomas at Huashan Hospital, Fudan University, between February 2017 and July 2019 were retrospectively analyzed. Two regions of interest (ROIs), called the maximum anomaly region (ROI1) and the tumor region (ROI2), were delineated on the patients' preoperative MRIs utilizing the tool ITK-SNAP, and Pyradiomics 3.0 was applied to execute feature extraction. Feature selection was performed utilizing a least absolute shrinkage and selection operator (LASSO) filter. Six classifiers, including Gaussian naive Bayes (GNB), random forest (RF), K-nearest neighbor (KNN), support vector machine (SVM) with a linear kernel, adaptive boosting (AB), and multilayer perceptron (MLP) were used to establish predictive models, and the predictive performance of the six classifiers was evaluated through five-fold cross-validation. The performance of the predictive models was evaluated using the AUC and other metrics. After that, the model with the best predictive performance was tested using the external data from The Cancer Imaging Archive (TCIA). According to the inclusion and exclusion criteria, 240 patients with gliomas were identified for inclusion in the study, including 106 grade 2, 68 grade 3, and 66 grade 4 gliomas. A total of 150 features was selected, and the MLP classifier had the best predictive performance among the six classifiers based on T2-FLAIR (mean AUC of 0.80 ± 0.07). The SVM classifier had the best predictive performance among the six classifiers based on DWI (mean AUC of 0.84 ± 0.05); the SVM classifier had the best predictive performance among the six classifiers based on CE-T1WI (mean AUC of 0.85 ± 0.06). Among the six classifiers, based on ROI1, the MLP classifier had the best prediction performance (mean AUC of 0.78 ± 0.07); among the six classifiers, based on ROI2, the SVM classifier had the best prediction performance (mean AUC of 0.82 ± 0.07). Among the six classifiers, based on the multimodal MRI of all the ROIs, the SVM classifier had the best prediction performance (average AUC of 0.85 ± 0.04). The SVM classifier, based on the multimodal MRI of all the ROIs, achieved an AUC of 0.81 using the external data from TCIA. The prediction model, based on preoperative conventional multimodal MRI radiomics, established in this study can conveniently, accurately, and noninvasively classify adult gliomas into grades 2-4, providing certain assistance for the precise diagnosis and treatment of patients and optimizing their clinical management.

Although numerous population pharmacokinetic (popPK) models of voriconazole have been developed, their predictive performance and extrapolation capacity remain uncertain. This study aimed to systematically evaluate the predictive performance of previously published voriconazole popPK models using an independent external clinical dataset, and to explore potential factors influencing model performance. Additionally, the study sought to identify candidate models with favorable extrapolation capability and clinical utility. The external evaluation dataset was derived from critically ill patients receiving extracorporeal membrane oxygenation (ECMO) support. Published voriconazole popPK models were selected and reconstructed. Predictive performance was assessed using both prediction- and simulation-based diagnostic tools. Furthermore, Bayesian forecasting was applied to evaluate the impact of prior information on the model's ability to predict individual plasma concentrations. A total of 76 voriconazole plasma concentrations from 19 patients were included, and 14 published models were included for evaluation. Diagnostic analyses revealed that none of the models fully satisfied the predefined thresholds for clinical acceptability. Both visual predictive checks (VPC) and normalized prediction distribution errors (NPDE) demonstrated varying degrees of predictive bias across all models, with individual models showing adequate performance only in specific diagnostic aspects. Bayesian forecasting indicated that increasing the number of prior concentrations did not consistently enhance predictive performance. In most models, the use of a single prior concentration was sufficient to markedly improve predictive accuracy and precision, achieving optimal performance. Notably, models incorporating allometric scaling exhibited the greatest improvement in Bayesian predictive performance. Whether the time interval between the prior and predicted concentrations affects predictive performance requires further investigation. Published voriconazole popPK models exhibited considerable variability in predictive performance. Multiple factors, including covariates, patient characteristics, and ethnic variability, may serve as critical determinants of model predictive performance and extrapolation capability. Integrating popPK modeling with maximum a posteriori Bayesian (MAPB) estimation and prior concentration information may serve as an effective approach to achieving model-informed precision dosing (MIPD) for voriconazole. Furthermore, incorporating allometric scaling into models may further enhance the performance of Bayesian prediction.

Real-time applications require simple techniques that enable predictable high performance rendering, but illumination methods for GPUs tend either to be complex and fragile on the one hand or very limited and offering poor visual quality on the other. The addition of a plausible global look to an application's lighting impacts users' perceptions of its realism, and on the spectrum of existing approaches, screen space ambient occlusion (SSAO) lies at the simplest and most predictable end. We have extended screen space ambient occlusion by replacing its depth buffer comparisons with a sampling of a volumetric discretization of the scene, while still evaluating it in an image-space post-process and thus retaining its predictable performance. We show improvements in quality over depth buffer based alternatives at a reasonable additional cost.

Background and objectiveDiabetes mellitus manifests as prolonged elevated blood glucose levels resulting from impaired insulin production. Such high glucose levels over a long period of time damage multiple internal organs. To mitigate this condition, researchers and engineers have developed the closed loop artificial pancreas consisting of a continuous glucose monitor and an insulin pump connected via a microcontroller or smartphone. A problem, however, is how to accurately predict short term future glucose levels in order to exert efficient glucose-level control. Much work in the literature focuses on least prediction error as a key metric and therefore pursues complex prediction methods such a deep learning. Such an approach neglects other important and significant design issues such as method complexity (impacting interpretability and safety), hardware requirements for low-power devices such as the insulin pump, the required amount of input data for training (potentially rendering the method infeasible for new patients), and the fact that very small improvements in accuracy may not have significant clinical benefit. MethodsWe propose a novel low-complexity, explainable blood glucose prediction method derived from the Intel P6 branch predictor algorithm. We use Meta-Differential Evolution to determine predictor parameters on training data splits of the benchmark datasets we use. A comparison is made between our new algorithm and a state-of-the-art deep-learning method for blood glucose level prediction. ResultsTo evaluate the new method, the Blood Glucose Level Prediction Challenge benchmark dataset is utilised. On the official test data split after training, the state-of-the-art deep learning method predicted glucose levels 30 min ahead of current time with 96.3% of predicted glucose levels having relative error less than 30% (which is equivalent to the safe zone of the Surveillance Error Grid). Our simpler, interpretable approach prolonged the prediction horizon by another 5 min with 95.8% of predicted glucose levels of all patients having relative error less than 30%. ConclusionsWhen considering predictive performance as assessed using the Blood Glucose Level Prediction Challenge benchmark dataset and Surveillance Error Grid metrics, we found that the new algorithm delivered comparable predictive accuracy performance, while operating only on the glucose-level signal with considerably less computational complexity.

PurposeTo develop and interpret optimal predictive models to identify epidermal growth factor receptor (EGFR) mutation status and subtypes in patients with lung adenocarcinoma based on multicentric 18F-FDG PET/CT data, and further construct a prognostic model to predict their clinical outcome.MethodsThe 18F-FDG PET/CT imaging and clinical characters of 767 patients with lung adenocarcinoma from 4 cohorts were collected. Seventy-six radiomics candidates using cross-combination method to identity EGFR mutation status and subtypes were built. Further, Shapley additive explanations and local interpretable model-agnostic explanations were used for optimal models’ interpretation. Moreover, in order to predict the overall survival, a multivariate Cox proportional hazard model based on handcrafted radiomics features and clinical characteristics was constructed. The predictive performance and clinical net benefit of the models were evaluated via area under receiver operating characteristic (AUC), C-index and decision curve analysis. ResultsAmong the 76 radiomics candidates, light gradient boosting machine classifier (LGBM) combined with recursive feature elimination wrapped LGBM feature selection method achieved best performance in predicting EGFR mutation status (AUC reached 0.80, 0.61, 0.71 in the internal test cohort and two external test cohorts, respectively). And extreme gradient boosting classifier combined with support vector machine feature selection method achieved best performance in predicting EGFR subtypes (AUC reached 0.76, 0.63, 0.61 in the internal test cohort and two external test cohorts, respectively). The C-index of the Cox proportional hazard model achieved 0.863.ConclusionsThe integration of cross-combination method and the external validation from multi-center data achieved a good prediction and generalization performance in predicting EGFR mutation status and its subtypes. The combination of handcrafted radiomics features and clinical factors achieved good performance in predicting prognosis. With the urgent needs of multicentric 18F-FDG PET/CT trails, robust and explainable radiomics models have great potential in decision making and prognosis prediction of lung adenocarcinoma.

To explore the key factors influencing hearing recovery in patients with sudden sensorineural hearing loss (SSNHL) and provide a basis for clinical prognosis evaluation and individualized treatment. Clinical data of 164 inpatients with SSNHL at a tertiary hospital in Fujian Province from January 2022 to December 2024 were retrospectively analyzed, including basic characteristics, hearing features, examination findings, and treatment measures. Univariate analyses (Chi-square test, Kruskal-Wallis H test) and a multivariate logistic regression model were used to identify prognostic factors, and a neural network model was constructed to evaluate its predictive performance. Among the 164 patients, 27 (16.46%) showed no response, 61 (37.20%) responded, 32 (19.51%) showed marked response, and 44 (26.83%) were cured. Univariate analysis revealed that age, length of hospital stay, degree of hearing loss, type of hearing loss, brain MRI findings, triglyceride (TG) levels, low-density lipoprotein (LDL-C) levels, and the use of glucocorticoids and batroxobin were associated with treatment outcomes (all P < .05). Multivariate logistic regression indicated that TG (OR = 0.559, P = .002) and LDL-C (OR = 0.352, P = .001) were independent risk factors, while glucocorticoid use was a protective factor (OR = 8.564, P = .004). The neural network model exhibited good predictive performance (AUC = 0.889, accuracy rate 86.5%). Advanced age, severe hearing loss (especially total deafness), abnormal brain MRI findings, and elevated TG and LDL-C levels may lead to poor prognosis in SSNHL, while early use of glucocorticoids can improve outcomes. The neural network model can effectively evaluate patients' potential for hearing recovery.

SummaryIn this paper, we develop a new model reference control architecture to effectively suppress system uncertainties and achieve a guaranteed transient and steady‐state system performance. Unlike traditional robust control frameworks, only a parameterization of the system uncertainty given by unknown weights with known conservative bounds is needed to stabilize uncertain dynamical systems with predictable system performance. In addition, the proposed architecture's performance is not dependent on the level of conservatism of the bounds of system uncertainty. Following the same train of thought as adaptive controllers that modify a given reference system to improve system performance, the proposed method is inspired by a recently developed command governor theory that minimizes the effect of system uncertainty by augmenting the input signal of the uncertain dynamical and reference systems. Specifically, a dynamical system, called a command governor, is designed such that its output is used to modify the input of both the controlled uncertain dynamical and reference systems. It is theoretically shown that if the command governor design parameter is judiciously selected, then the controlled system approximates the given original, unmodified reference system. The proposed approach is advantageous over model reference adaptive control approaches because linearity of the uncertain dynamical system is preserved through linear control laws, and hence, the closed‐loop performance is predictable for different command spectrums. Additionally, it is shown that the architecture can be modified for robustness improvements with respect to high frequency content due to, for example, measurement noise. Modifications can also be made in order to accommodate actuator dynamics and retain closed‐loop stability and predictable performance. The main contribution of this paper is the rigorous analysis of the stability and performance of a system utilizing the command governor framework. A numerical example is provided to illustrate the effectiveness of the proposed architecture. Copyright © 2015 John Wiley &amp; Sons, Ltd.

The dengue virus is the most dangerous one that mosquitoes may spread to people. Despite attempts by the government, dengue outbreaks are becoming increasingly common in Bangladesh. Interventions in public health rely heavily on KAP (knowledge, attitude, and practice) studies. The primary goal of this research is to forecast the occurrence of dengue disease in the city of Dhaka using methods from machine learning and time series analysis and then to compare the models in order to find the one with the lowest MAPE. From January 2016 through July 2021, monthly data were retrieved for this study from WHO and the Directorate General of Health and Services (DGHS). According to the findings of this research, neural networks outperform time series analysis when it comes to making predictions. The best‐fitted neural network (NN) model was found in model 04 with 05 hidden layers which produced the minimum error model with the value of error 0.003032557, and the values of RMSE and MAPE are 7.588889e − 06 and 1.15273, respectively, for the prediction of the dengue fever in Dhaka city. In contrast, the original dengue data in the time series analysis is not stationary. Take the difference and run the unit root test by the augmented Dickey–Fuller (ADF) test to make it stationary. The dengue data series is stationary at the first‐order difference, as evidenced by the ACF and PACF, which show no noticeable spike in the first‐order difference. The ARIMA (6, 1, 1) model with the lowest AIC = −251.8, RMSE = 0.0310797, and MAPE = 15.2892 is the best choice model for predicting the dengue death rate. Therefore, from these two models, the NN model gives better prediction performance with the lowest value of MAPE. So, the neural network gives better prediction performance than time series analysis. The NN model forecasted 12‐month death rates of dengue fever that suggest the death rate in dengue fever falling month by month. This study is more innovative than any other research because this research approach is different from any other research approach. The model selection criteria are based on the most effective performance metrics MAPE, indicating the lowest error and better prediction performance. Therefore, from this research, the author suggests machine learning gives better prediction performance than time series analysis for any other prediction performance.

Purpose: This study assessed and demonstrated the efficacy of two select empirical forecast models (i.e., autoregressive integrated moving average [ARIMA] model vs. grey model [GM]) in accurately predicting state vocational rehabilitation agency (SVRA) rehabilitation success rate trends across six different racial and ethnic population cohorts (i.e., Blacks or African Americans, non-Latino Whites, American Indians or Alaskan Natives, Asians, Native Hawaiians or other Pacific Islanders, and Latinos). Methods: Eleven years of Rehabilitation Services Administration (RSA)-911 case record data (fiscal year [FY] 2004 through 2014) on SVRA employment outcomes were extracted and entered into the ARIMA model and GM to test their predictive performance. Results: The GM was demonstrated to be superior to the ARIMA model in predictive accuracy performance. Remarkably, although the GM (1, 1) 3-year frequency curve projection simulation results (FY 2015–2017) showed a slight upward trajectory in the number of successfully rehabilitated Latinos compared to baseline FY 2014 actual numbers, more drastic downward trajectories were projected for Blacks or African Americans, non-Latino Whites, American Indians or Alaskan Natives, Asians, and Native Hawaiians or other Pacific Islanders. Conclusions: The GM represents a demonstrably capable and promising forecasting tool that could be useful to SVRA leaders, policy makers, advocates, and researchers in simulating predictions that inform future policy initiatives, influence strategic plan development, and help guide the state of the science on future research and development foci. Additional multiple comprehensive demonstration trials, nonetheless, are needed to either confirm or refute the GM’s veracity in national and state predictive accuracy and curve fitting performance.