Search

The melting temperature is important for materials design because of its relationship with thermal stability, synthesis, and processing conditions. Current empirical and computational melting point estimation techniques are limited in scope, computational feasibility, or interpretability. We report the development of a machine learning methodology for predicting melting temperatures of binary ionic solid materials. We evaluated different machine-learning models trained on a dataset of the melting points of 476 non-metallic crystalline binary compounds using materials embeddings constructed from elemental properties and density-functional theory calculations as model inputs. A direct supervised-learning approach yields a mean absolute error of around 180 K but suffers from low interpretability. We find that the fidelity of predictions can further be improved by introducing an additional unsupervised-learning step that first classifies the materials before the melting-point regression. Not only does this two-step model exhibit improved accuracy, but the approach also provides a level of interpretability with insights into feature importance and different types of melting that depend on the specific atomic bonding inside a material. Motivated by this finding, we used a symbolic learning approach to find interpretable physical models for the melting temperature, which recovered the best-performing features from both prior models and provided additional interpretability. [ABSTRACT FROM AUTHOR]

Nanoconfined water plays an important role in broad fields of science and engineering. Classical molecular dynamics (MD) simulations have been widely used to investigate water phases under nanoconfinement. The key ingredient of MD is the force field. In this study, we systematically investigated the performance of a recently introduced family of globally optimal water models, OPC and OPC3, and TIP4P/2005 in describing nanoconfined two-dimensional (2D) water ice. Our studies show that the melting points of the monolayer square ice (MSI) of all three water models are higher than the melting points of the corresponding bulk ice Ih. Under the same conditions, the melting points of MSI of OPC and TIP4P/2005 are the same and are ∼90 K lower than that of the OPC3 water model. In addition, we show that OPC and TIP4P/2005 water models are able to form a bilayer AA-stacked structure and a trilayer AAA-stacked structure, which are not the cases for the OPC3 model. Considering the available experimental data and first-principles simulations, we consider the OPC water model as a potential water model for 2D water ice MD studies. [ABSTRACT FROM AUTHOR]

Solar interfacial evaporation, as a novel practical freshwater production method, requires continuous research on how to improve the evaporation rates to increase water production. In this study, sets of data were obtained from molecule dynamics simulation and literature, in which the parameters included height, diameter, height–radius ratio, evaporation efficiency, and evaporation rate. Initially, the correlation between the four input parameters and the output of the evaporation rate was examined through traditional pairwise plots and Pearson correlation analysis, revealing weak correlations. Subsequently, the accuracy and generalization performance of the evaporation rate prediction models established by neural network and random forest were compared, with the latter demonstrating superior performance and reliability confirmed via random data extraction. Furthermore, the impact of different percentages (10%, 20%, and 30%) of the data on the model performance was explored, and the result indicated that the model performance is better when the test set is 20% and all the constructed model converge. Moreover, the mean absolute error and mean squared error of the evaporation rate prediction model for the three ratios were calculated to evaluate their performance. However, the relationship between the height- radius ratio and optimal evaporation rate was investigated using the enumeration method, and it was determined that the evaporation efficiency was optimal when the height–radius ratio was 6. Finally, the importance of height, diameter, height– radius ratio, and evaporation efficiency were calculated to optimize evaporator structure, increase evaporation rate, and facilitate the application of interfacial evaporation in solar desalination. [ABSTRACT FROM AUTHOR]

Multi-scale predictive models for the shock sensitivity of energetic materials connect energy localization ("hotspots") in the microstructure to macro-scale detonation phenomena. Calculations of hotspot ignition and growth rely on models for chemical reaction rates expressed in Arrhenius forms; these chemical kinetic models, therefore, are foundational to the construction of physics-based, simulation-derived meso-informed closure (reactive burn) models. However, even for commonly used energetic materials (e.g., HMX in this paper) there are a wide variety of reaction rate models available. These available reaction rate models produce reaction time scales that vary by several orders of magnitude. From a multi-scale modeling standpoint, it is important to determine which model best represents the reactive response of the material. In this paper, we examine three global Arrhenius-form rate models that span the range of reaction time scales, namely, the Tarver 3-equation, the Henson 1-equation, and the Menikoff 1-equation models. They are employed in a meso-informed ignition and growth model which allows for connecting meso-scale hotspot dynamics to macro-scale shock-to-detonation transition. The ability of the three reaction models to reproduce experimentally observed sensitivity is assessed by comparing the predicted criticality envelope (Walker–Wasley curve) with experimental data for pressed HMX Class V microstructures. The results provide a guideline for model developers on the plausible range of time-to-ignition that are produced by physically correct Arrhenius rate models for HMX. [ABSTRACT FROM AUTHOR]

We present a theoretical model to study the effect of counterion size on the effective charge, size, and thermodynamic behavior of a single, isolated, and flexible polyelectrolyte (PE) chain. We analyze how altering counterion size modifies the energy and entropy contributions to the system, including the ion-pair free energy, excluded volume interactions, entropy of free and condensed ions, and dipolar attraction among monomer–counterion pairs, which result in competing effects challenging intuitive predictions. The PE self-energy is calculated using the Edwards–Muthukumar Hamiltonian, considering a Gaussian monomer distribution for the PE. The condensed ions are assumed to be confined within a cylindrical volume around the PE backbone. The dipolar and excluded volume interactions are described by the second and third virial coefficients. The assumption of freely rotating dipoles results in a first-order coil–globule transition of the PE chain. A more realistic, weaker dipolar attraction, parameterized in our theory, shifts it to a second-order continuous transition. We calculate the size scaling-exponent of the PE and find exponents according to the relative dominance of the electrostatic, excluded volume, or dipolar effects. We further identify the entropy- and energy-driven regimes of the effective charge and conformation of the PE, highlighting the interplay of free ion entropy and ion-pair energy with varying electrostatic strengths. The crossover strength, dependent on the counterion size, indicates that diminishing sizes favor counterion condensation at the expense of free ion entropy. The predictions of the model are consistent with trends in simulations and generalize the findings of the point-like counterion theories. [ABSTRACT FROM AUTHOR]

Portevin–Le Chatelier effect, associated with dynamic strain aging, is widely found in various alloys, and the maximum serration magnitude from the tensile stress–strain curve can be considered as the point closest to the failure threshold. Due to the difficulty to define solute and solvent atoms in multicomponent alloys, it is a great challenge to explore the serration mechanism. In this study, the atoms that play the role of pinning are defined as solute atoms. Aided by mean-field theory, an in situ pinning model coupled with the twinning effect in a CoNiV alloy with low stacking fault energies is successfully established, which can well predict the failure threshold, i.e., the maximum serration magnitude. The present study paves a new way to correlate the serration dynamics and in situ pinning, and further predicts the failure threshold upon loading for multicomponent high- and medium-entropy alloys. [ABSTRACT FROM AUTHOR]

Structural integrity and robustness are key parameters to evaluate microfabrication techniques. Bending and collapsing of 2D/3D microstructures are commonly noted in solvent-involved procedures, e.g., liquid-based post-treatment in wet-etching, lithography, and Two Photon Polymerizations (TPPs). Such structural failures are caused by excessive solution-imposed capillary forces, where multiple kinds of liquids may intensively participate. Current pieces of the literature focus on the mechanical one-axis models to illustrate their deformation process. To date, there exists an emerging demand for dual-axis models to satisfy rapidly developed micro/nano-engineerings. Here, utilizing polymer micro-pillars distributed in a square array as an illustration example, a dual-axis beam-sway model is proposed considering the influences of structure arrangement as well as the solvent. Specifically, a simplified criterion for judging structural stability is identified. For verifications, the TPP-based experimental data show excellent consistency with model predictions. All in all, the extended model offers reliable guidance for the fabrication of delicate structures and further benefits the optimization of related microfabrication processes. [ABSTRACT FROM AUTHOR]

A model predicting the number of emission sites and total current from a porous conical electrospray emitter as functions of voltage is derived. A pressure balance between capillary and electric forces is used to determine an onset criterion for individual menisci, and an ionic emission scaling law is invoked to predict the current each meniscus emits. These submodels are integrated over a phenomenological meniscus size distribution and the area of the emitter to yield a model for emitter performance as a function of five free parameters, two for the ionic emission submodel and three for the meniscus size distribution. Bayesian inference is applied to determine these model parameters from an existing dataset [Dressler et al., J. Propul. Power 38, 809 (2022)]. The model predictions after training are compared to the experimental data, and it is found that the majority of the data are within a 90% credible interval. The ability of the model to capture key trends in the experimental data is attributed to the interplay of two effects: the distribution over meniscus size on the emitter and the position-dependent electric field. The calibrated model results also suggest that the emitter surface is wetted by a series of large but sparsely distributed pools of propellant. The performance and extensibility of the model are examined within the context of model-based design for porous electrospray array thrusters. [ABSTRACT FROM AUTHOR]

BACKGROUND: The distribution of the necrotic area plays an important role in hip preservation treatment. At present, there are few studies on whether the difference in the three-dimensional spatial distribution of osteonecrosis of the femoral head affects the clinical outcome of fibular support. OBJECTIVE: To explore the relationship between the spatial distribution and clinical outcome at the sites of osteonecrosis of the femoral head and fibular support using CT three-dimensional reconstruction so as to provide a basis for optimizing the applicable conditions of fibular support and improving the hip preservation effect of fibular support. METHODS: Eighty patients with osteonecrosis of the femoral head who were treated with fibular support for hip preservation from January 2010 to January 2021 were selected as the study subjects according to the inclusion criteria. They were followed up for at least 2 years. According to the clinical outcome, the patients were divided into the successful hip preservation group (n=55) and the failure hip preservation group (n=25). 3D reconstruction was performed according to the preoperative and postoperative CT images of the patients. According to the three-column theory, the femoral head was divided into outer nine areas, middle nine areas and inner nine areas (L1-9, C1-9, and M1-9) to explore the spatial distribution of necrotic area of the femoral head and fibular support area and its relationship with clinical outcome. RESULTS AND CONCLUSION: (1) Before operation, the necrotic area of the femoral head was mainly distributed in L1, L2, L4, L5, C1, C2, C4, and C5 (the upper and middle part of the anterior part of the outer ninth area and the middle part of the middle ninth area). After operation, the fibular support area was mainly distributed in L5, L6, C5, and C6 (the middle and lower part of the outer ninth area and the middle and lower part of the middle ninth area). (2) There were significant differences in the distribution of osteonecrosis of the femoral head between the successful hip preservation group and the failure hip preservation group in L8 (the posterior middle part of the outer ninth area), C3 (the anterior lower part of the middle ninth area), C6 (the lower middle part of the middle part of the inner ninth area) and M2 (the anterior middle part of the inner ninth area) (P < 0.05). There was a significant difference in the distribution of fibular support in L5 and L6 (middle and lower part of outer nine) (P < 0.05). Among them, the L8 region could be used as an independent predictor of hip preservation failure in fibular support surgery. The area under the curve of the L8 single factor prediction model was 0.698 [95%CI (0.575, 0.822)]; the sensitivity was 76%, and the specificity was 63.6%. (3) It turns out, when the necrotic area involves L8, C3, C6, and M2, especially L8, the failure of fibular support may increase, and when the fibular support involves L5 and L6, the effect of hip preservation is often not ideal. [ABSTRACT FROM AUTHOR]

This article reports on development of an adaptive framework for predicting the erosion performance of polymer composites using certain statistical and machine learning (ML) models. For this, ramie‐epoxy composites reinforced with variations (0–30 wt%) of sponge iron slag (an iron industry waste) are considered. The composites are fabricated and then subjected to high temperature solid particle erosion wear trials following Taguchi's L27 orthogonal array. The effects of different control factors on the erosion rate in an interactive environment are appraised by analysis of variance (ANOVA) which reveals the filler content as the most significant factor contributing 66.21%, followed by impact velocity (22.86%) and impingement angle (2.28%). A regression model based on the input–output parameters obtained from experimentation is constructed for prediction of erosion rate. Further, four predictive models using different machine learning algorithms are also proposed to predict the erosion rate of the composites. The feasibility and performance of each ML model is assessed using appropriate performance metrics. Among all the models, the gradient boosting machine model is found to be the most reliable model exhibiting the highest prediction accuracy and least errors. Highlights: Development of novel class of composites reinforced with sponge iron slag.Database creation based on erosion wear experimentation on the composites.Data‐driven modeling for prediction of erosion rates using machine learning.Comparison of performance of different models and identifying the best one. [ABSTRACT FROM AUTHOR]

Despite documented associations between social determinants of health and outcomes post–congenital heart surgery, clinical risk models typically exclude these factors. The study sought to characterize associations between social determinants and operative and longitudinal mortality as well as assess impacts on risk model performance. Demographic and clinical data were obtained for all congenital heart surgeries (2006-2021) from locally held Congenital Heart Surgery Collaborative for Longitudinal Outcomes and Utilization of Resources Society of Thoracic Surgeons Congenital Heart Surgery Database data. Neighborhood-level American Community Survey and composite sociodemographic measures were linked by zip code. Model prediction, discrimination, and impact on quality assessment were assessed before and after inclusion of social determinants in models based on the 2020 Society of Thoracic Surgeons Congenital Heart Surgery Database Mortality Risk Model. Of 14,173 total index operations across New York State, 12,321 cases, representing 10,271 patients at 8 centers, had zip codes for linkage. A total of 327 (2.7%) patients died in the hospital or before 30 days, and 314 children died by December 31, 2021 (total n = 641; 6.2%). Multiple measures of social determinants of health explained as much or more variability in operative and longitudinal mortality than clinical comorbidities or prior cardiac surgery. Inclusion of social determinants minimally improved models' predictive performance (operative: 0.834-0.844; longitudinal 0.808-0.811), but significantly improved model discrimination; 10.0% more survivors and 4.8% more mortalities were appropriately risk classified with inclusion. Wide variation in reclassification was observed by site, resulting in changes in the center performance classification category for 2 of 8 centers. Although indiscriminate inclusion of social determinants in clinical risk modeling can conceal inequities, thoughtful consideration can help centers understand their performance across populations and guide efforts to improve health equity. [Display omitted] [ABSTRACT FROM AUTHOR]

Tensile testing (aka tension testing) is a widely employed mechanical testing technique for analyzing materials' properties and behavior under applied stress. Tensile testing plays a pivotal role in helping engineers to make informed decision about material selection and usage. Despite its importance, there is a limited numbers of studies that explored the potential of AI techniques for real time monitoring and material behavior prediction in tensile testing. To this end, this work presents a deep learning model designed to predict the material's condition throughout tensile testing and provide an early warning prior to fracture. By leveraging a comprehensive dataset of tension test video samples, the proposed model utilizes both convolution and recurrent neural networks to extract pertinent spatial and temporal visual features, thereby predicting the frames at which material deformation and fracture occur. The evaluation results of our research showed that the proposed model achieved a predictive ability with an F1-score of 97%, on average. The implications of our research are significant for industries and researchers in the field of materials science and engineering. By accurately predicting material status, our model enables autonomous, real time analysis of material behavior during tensile testing, leading to better time and cost efficiency in various applications. [ABSTRACT FROM AUTHOR]

To achieve smart and sustainable construction goals, machine learning (ML) techniques can serve as a cost-effective and efficient substitute for labour-intensive, laboratory, or in situ approaches in parameter estimation essential for infrastructure design. Of these, soil's shear strength parameters, notably cohesion (c) and friction angle (φ), typically govern the design of geo-structures. For quick and cost-effective estimation of these parameters, the earlier studies proposed ML-based predictive models that were less practical and accurate or consider an excessive number of input variables. To minimize these limitations, our study proposes new models of c and φ using gene expression programming (GEP) based on readily available soil attributes such as sand content (S), depth (D), specific gravity (Gs), liquid limit (LL), plastic limit (PL), and fine content (FC). The newly proposed models show excellent accuracy as the values of R2, RMSE (root mean square error), MAE (mean absolute error), RSE (relative standard error) for c-predictive model were 0.984, 1.13, 0.878, 0.017, respectively, and were 0.927, 1.123, 0.922, 0.072, respectively, for φ-predictive model. Through sensitivity analysis, FC and LL emerged as the most critical parameters influencing c, while Gs and PL proved sensitive for determining φ. In comparison with existing models, the c-predictive model displays R2 enhancements of 11.84–45.87% and RMSE improvements of 65.9–92.03%, while the φ-predictive model showcases R2 gains of 13.16–23.75% and RMSE improvements of 58.79–69.29%. By integrating predictive prowess with sustainable and smart construction principles, our study plots a realistic course for efficient geo-structural design. [ABSTRACT FROM AUTHOR]

With the plentiful information available in the Gaia BP/RP spectra, there is significant scope for applying discriminative models to extract stellar atmospheric parameters and abundances. We describe an approach to leverage an 'Uncertain Neural Network' model trained on APOGEE data to provide high-quality predictions with robust estimates for per-prediction uncertainty. We report median formal uncertainties of 0.068 dex, 69.1 K, 0.14 dex, 0.031 dex, 0.040 dex, and 0.029 dex for [Fe/H], T eff, log  g , [C/Fe], [N/Fe], and [α/M], respectively. We validate these predictions against our APOGEE training data, LAMOST, and Gaia GSP- phot stellar parameters, and see a strong correlation between our predicted parameters and those derived from these surveys. We investigate the information content of the spectra by considering the 'attention' our model pays to different spectral features compared to expectations from synthetic spectra calculations. Our model's predictions are applied to the Gaia data set, and we produce a publicly available catalogue of our model's predictions. [ABSTRACT FROM AUTHOR]

The COVID-19 pandemic has profoundly impacted healthcare systems and economies worldwide, leading to the implementation of travel restrictions and social measures. Efforts such as vaccination campaigns, testing, and surveillance have played a crucial role in containing the spread of the virus and safeguarding public health. There needs to be more research exploring the transmission dynamics of COVID-19, particularly within European nations. Therefore, the primary objective of this research was to examine the spread patterns of COVID-19 across various European countries. Doing so makes it possible to implement preventive measures, allocate resources, and optimize treatment strategies based on projected case and mortality rates. For this purpose, a hybrid prediction model combining CNN and LSTM models was developed. The performance of this hybrid model was compared against several other models, including CNN, k-NN, LR, LSTM, MLP, RF, SVM, and XGBoost. The empirical findings revealed that the CNN-LSTM hybrid model exhibited superior performance compared to alternative models in effectively predicting the transmission of COVID-19 within European nations. Furthermore, examining the peak of case and death dates provided insights into the dynamics of COVID-19 transmission among European countries. Chord diagrams were drawn to analyze the inter-country transmission patterns of COVID-19 over 5-day and 14-day intervals. [ABSTRACT FROM AUTHOR]

This article investigates an observer-based robust model predictive control (RMPC) design to control the uncertain discrete-time linear systems with disturbances. To make a more practical scheme, it is supposed that the uncertain system has been faced with unknown disturbance and input constraints. The proposed RMPC approach is based on a state feedback control design that ensures the H ∞ performance criterion to attenuate disturbance affections. Furthermore, in view of practical application, the control law is constructed based on the estimated states obtained from the Luenberger observer. Based on Lyapunov's theory, the input to state practically stability (ISPS) of the closed-loop system is ensured. Appropriate conditions for the ISPS of the closed-loop system and the estimation error are obtained in terms of online linear matrix inequalities (LMIs) which lead to obtaining the time-varying gain matrices of both controller and observer. Finally, to validate the obtained results, the proposed approach is applied to a numerical example and it is compared with the existing control scheme and the superiority is proved. [ABSTRACT FROM AUTHOR]

Background: Hepatocellular carcinoma (HCC) is a prevalent malignancy with a pressing need for improved therapeutic response and prognosis prediction. This study delves into a novel predictive model related to ferroptosis, a regulated cell death mechanism disrupting metabolic processes. Results: Single-cell sequencing data analysis identified subpopulations of HCC cells exhibiting activated ferroptosis and distinct gene expression patterns compared to normal tissues. Utilizing the LASSO-Cox algorithm, we constructed a model with 10 single-cell biomarkers associated with ferroptosis, namely STMN1, S100A10, FABP5, CAPG, RGCC, ENO1, ANXA5, UTRN, CXCR3, and ITM2A. Comprehensive analyses using these biomarkers revealed variations in immune infiltration, tumor mutation burden, drug sensitivity, and biological functional profiles between risk groups. Specific associations were established between particular immune cell subtypes and certain gene expression patterns. Treatment response analyses indicated potential benefits from anti-tumor immune therapy for the low-risk group and chemotherapy advantages for the high-risk group. Conclusions: The integration of this single-cell level model with clinicopathological features enabled accurate overall survival prediction and effective risk stratification in HCC patients. Our findings illuminate the potential of ferroptosis-related genes in tailoring therapy and prognosis prediction for HCC, offering novel insights into the intricate interplay among ferroptosis, immune response, and HCC progression. [ABSTRACT FROM AUTHOR]

The thermal management of proton exchange membrane fuel cell crucially impacts the performance and durability of fuel cell vehicle systems. Optimized temperature control is critical to the water balance and electrochemical reaction inside the fuel cell stack under dynamic operating conditions. This work develops a dynamic thermal management system model with uncertain parameters for the dual coolant circulation fuel cell system to simulate unforeseeable disturbances in real systems. With the proposed model, the tube-based robust model predictive control (tube-RMPC) is innovatively applied to optimize dynamic thermal management. This strategy can identify the optimal trajectories of thermostat angle, heater power, water pump speed, and cooling fan speed, thus maximizing the temperature dynamic response speed while minimizing the stack temperature difference fluctuations in the presence of unforeseeable disturbances. The result shows that during the load ascending and descending operations, the temperature dynamic response time is 8 and 15 s, with temperature fluctuations below 1 °C. The root mean square error for stack temperature is reduced by 10.4% and 24.4% compared to the closed-loop MPC. The fuel cell system exhibits excellent temperature dynamic response capability using the tube-RMPC method. • Tube-RMPC is applied in fuel cell vehicle systems to effectively suppress stack temperature fluctuations. • Robust dynamic thermal management strategy is proposed to determine optimal action trajectories of critical TMS components. • Proposed method controls temperatures at different stack locations precisely under rapid load changes. • Robustness of tube based MPC under unpredictable disturbances is validated. [ABSTRACT FROM AUTHOR]

Background: Diabetic kidney disease (DKD) is the main cause of end-stage renal disease and has a high mortality rate. Currently, no effective treatments are available to reduce the progression of kidney damage associated with diabetes. Objectives: To explore the influence and predictive value of the atherogenic index of plasma (AIP) on early chronic kidney disease and liver injury in patients with type 2 diabetes mellitus (T2DM). Methods: Medical records of 1057 hospitalized adult patients with T2DM between January 2021 and December 2022 were collected. The predictive value of AIP, renal function, and liver injury in patients with T2DM were analyzed using Pearson's correlation, multiple logistic regression, and receiver operating characteristic (ROC) curve analyses. Results: AIP was a sensitive indicator of early liver and kidney injury in patients with T2DM. Patients in the DKD group showed increased AIP that positively correlated with serum creatinine, uric acid, and β2-microglobulin levels. Increased AIP negatively correlated with estimated glomerular filtration rate (eGFR). AIP significantly correlated with alanine aminotransferase and aspartate aminotransferase levels and glutamyl transpeptidase-to-platelet ratio (GPR). An eGFR of 60-100 mL/min/1.73 m2 significantly increased the risk of DKD as the AIP increased. At lower GPR levels, the risk of DKD significantly increased with increasing AIP. However, no significant difference was found between the 2 groups when the GPR was >0.1407. The ROC curve analysis showed that AIP could predict early liver injury. Conclusions: AIP is directly involved in early liver and kidney injury in T2DM and may be a sensitive indicator for early detection. Plain Language Summary: Diabetes and its complications are a global public health concern. Diabetic kidney disease (DKD) is the main cause of end-stage renal disease, and metabolism-related disease factors are found throughout the progression of DKD. This study identified common sensitive indicators of early metabolism-related damage to liver and kidney function in patients with T2DM. [ABSTRACT FROM AUTHOR]

Background: Postoperative urinary retention (POUR), a common condition after prolapse surgery with potential serious sequelae if left untreated, lacks a clearly established optimal timing for catheter removal. This study aimed to develop and validate a predictive model for postoperative urinary retention lasting > 2 and > 4 days after prolapse surgery. Methods: We conducted a retrospective review of 1,122 patients undergoing prolapse surgery. The dataset was divided into training and testing cohorts. POUR was defined as the need for continuous intermittent catheterization resulting from a failed spontaneous voiding trial, with passing defined as two consecutive voids ≥ 150 mL and a postvoid residual urine volume ≤ 150 mL. We performed logistic regression and the predicted model was validated using both training and testing cohorts. Results: Among patients, 31% and 12% experienced POUR lasting > 2 and > 4 days, respectively. Multivariable logistic model identified 6 predictors. For predicting POUR, internal validation using cross-validation approach showed good performance, with accuracy lasting > 2 (area under the curve [AUC] 0.73) and > 4 days (AUC 0.75). Split validation using pre-separated dataset also showed good performance, with accuracy lasting > 2 (AUC 0.73) and > 4 days (AUC 0.74). Calibration curves demonstrated that the model accurately predicted POUR lasting > 2 and > 4 days (from 0 to 80%). Conclusions: The proposed prediction model can assist clinicians in personalizing postoperative bladder care for patients undergoing prolapse surgery by providing accurate individual risk estimates. [ABSTRACT FROM AUTHOR]

CRISPR is a gene editing technology which enables precise in-vivo genome editing; but its potential is hampered by its relatively low specificity and sensitivity. Improving CRISPR's on-target and off-target effects requires a better understanding of its mechanism and determinants. Here we demonstrate, for the first time, the chromosomal 3D spatial structure's association with CRISPR's cleavage efficiency, and its predictive capabilities. We used high-resolution Hi-C data to estimate the 3D distance between different regions in the human genome and utilized these spatial properties to generate 3D-based features, characterizing each region's density. We evaluated these features based on empirical, in-vivo CRISPR efficiency data and compared them to 425 features used in state-of-the-art models. The 3D features ranked in the top 13% of the features, and significantly improved the predictive power of LASSO and xgboost models trained with these features. The features indicated that sites with lower spatial density demonstrated higher efficiency. Understanding how CRISPR is affected by the 3D DNA structure provides insight into CRISPR's mechanism in general and improves our ability to correctly predict CRISPR's cleavage as well as design sgRNAs for therapeutic and scientific use. Author summary: CRISPR is an efficient and precise genome editing technology; but our understanding of the factors influencing it, and our ability to accurately design and predict its efficiency at different target sites, are lacking. Current models mostly use the sequence of the target site, rather than incorporating additional information regarding its genomic location. Here we propose a new type of predictive feature, based on the 3D structure of the genome. We show that features estimating the spatial density of the target site correlate highly with CRISPR efficiency, compared to classic sequence-based features; and that adding these features to predictive models significantly improves their power. CRISPR efficiency was negatively correlated with 3D density, indicating CRISPR is more efficient in sparse regions–possibly due to easier access to the target site. Improving our ability to predict CRISPR action will allow us to further understand its mechanism and better utilize it in research and medicine. [ABSTRACT FROM AUTHOR]

The key to extractive distillation separation lies in screening a suitable entrainer. Based on chemical process system engineering (PSE) concepts, a novel heuristic process prediction model is proposed with a minimized process economic cost index (PECI) for the efficient screening of an optimal green entrainer with the advantages of cost-effectiveness, environmental friendliness, and sustainability, followed by outlining the principles of characterizations and mathematical modeling in chemical ED processes. The relationships of αHC/SOL and NTR with QR, TAC, PECI, and LTEDI indicated that the heuristic process prediction model was scientifically valid as well as efficient. Moreover, the proposed model progresses from binomial intersecting influencing factors to trinomial juxtaposing influencing factors; summarizes and discloses the system’s theoretical rules and influencing elements of the changes in parameters and their relative volatility, avoiding the new azeotropes in the solvent recovery column; reduces the workloads of process simulation and optimization of distillation processes from (M × N) times to (1 + N) times; can be extended efficiently and accurately to provide the targeted prediction and systematic theoretical principles for separating different feed azeotropes, predicting and screening organic solvents, ionic liquids, deep eutectic solvents, and mixed solvents with the optimal TAC and environmental friendliness (AP and GWP) and sustainability (FETP, HTP, and TETP) as LCA impacts; has broad predictability and applicability combined with different process optimization algorithms; and provides infinite composite patterns and opportunities for integrating and developing innovative optimization module algorithms for future green chemical separation processes. [ABSTRACT FROM AUTHOR]

The selection of highly productive genotypes with stable performance across environments is a major challenge of plant breeding programs due to genotype-by-environment (GE) interactions. Over the years, different metrics have been proposed that aim at characterizing the superiority and/or stability of genotype performance across environments. However, these metrics are traditionally estimated using phenotypic values only and are not well suited to an unbalanced design in which genotypes are not observed in all environments. The objective of this research was to propose and evaluate new estimators of the following GE metrics: Ecovalence, Environmental Variance, Finlay–Wilkinson regression coefficient, and Lin–Binns superiority measure. Drawing from a multi-environment genomic prediction model, we derived the best linear unbiased prediction for each GE metric. These derivations included both a squared expectation and a variance term. To assess the effectiveness of our new estimators, we conducted simulations that varied in traits and environment parameters. In our results, new estimators consistently outperformed traditional phenotype-based estimators in terms of accuracy. By incorporating a variance term into our new estimators, in addition to the squared expectation term, we were able to improve the precision of our estimates, particularly for Ecovalence in situations where heritability was low and/or sparseness was high. All methods are implemented in a new R-package: GEmetrics. These genomic-based estimators enable estimating GE metrics in unbalanced designs and predicting GE metrics for new genotypes, which should help improve the selection efficiency of high-performance and stable genotypes across environments. [ABSTRACT FROM AUTHOR]

Background: Cervical spondylotic myelopathy (CSM) is a prevalent degenerative condition resulting from spinal cord compression and injury. Laminectomy with posterior spinal fusion (LPSF) is a commonly employed treatment approach for CSM patients. This study aimed to assess the effectiveness of machine learning models (MLMs) in predicting clinical outcomes in CSM patients undergoing LPSF. Methods: A retrospective analysis was conducted on 329 CSM patients who underwent LPSF at our institution from Jul 2017 to Jul 2023. Neurological outcomes were evaluated using the modified Japanese Orthopaedic Association (mJOA) scale preoperatively and at the final follow-up. Patients were categorized into two groups based on clinical outcomes: the favorable group (recovery rates ≥ 52.8%) and the unfavorable group (recovery rates < 52.8%). Potential predictors for poor clinical outcomes were compared between the groups. Four MLMs—random forest (RF), logistic regression (LR), support vector machine (SVM), and k-nearest neighborhood (k-NN)—were utilized to predict clinical outcome. RF model was also employed to identify factors associated with poor clinical outcome. Results: Out of the 329 patients, 185 were male (56.2%) and 144 were female (43.4%), with an average follow-up period of 17.86 ± 1.74 months. Among them, 267 patients (81.2%) had favorable clinical outcomes, while 62 patients (18.8%) did not achieve favorable results. Analysis using binary logistic regression indicated that age, preoperative mJOA scale, and symptom duration (p < 0.05) were independent predictors of unfavorable clinical outcomes. All models performed satisfactorily, with RF achieving the highest accuracy of 0.922. RF also displayed superior sensitivity and specificity (sensitivity = 0.851, specificity = 0.944). The Area under the Curve (AUC) values for RF, Logistic LR, SVM, and k-NN were 0.905, 0.827, 0.851, and 0.883, respectively. The RF model identified preoperative mJOA scale, age, symptom duration, and MRI signal changes as the most significant variables associated with poor clinical outcomes in descending order. Conclusions: This study highlighted the effectiveness of machine learning models in predicting the clinical outcomes of CSM patients undergoing LPSF. These models have the potential to forecast clinical outcomes in this patient population, providing valuable prognostic insights for preoperative counseling and postoperative management. [ABSTRACT FROM AUTHOR]

Objective: To compare the predictive efficacy of the PADUA and Caprini models for pulmonary embolism (PE) in gynecological inpatients, analyze the risk factors for PE, and validate whether both models can effectively predict mortality rates. Methods: A total of 355 gynecological inpatients who underwent computed tomography pulmonary angiography (CTPA) were included in the retrospective analysis. The comparative assessment of the predictive capabilities for PE between the PADUA and Caprini was carried out using receiver operating characteristic (ROC) curves. Logistic regression analysis was used to identify risk factors associated with PE. Additionally, Kaplan–Meier survival analysis plots were generated to validate the predictive efficacy for mortality rates. Results: Among 355 patients, the PADUA and Caprini demonstrated the area under the curve (AUC) values of 0.757 and 0.756, respectively. There was no statistically significant difference in the AUC between the two models (P = 0.9542). Multivariate logistic analysis revealed immobility (P < 0.001), history of venous thromboembolism (VTE) (P = 0.002), thrombophilia (P < 0.001), hormonal treatment (P = 0.022), and obesity (P = 0.019) as independent risk factors for PE. Kaplan–Meier survival analysis demonstrated the reliable predictive efficacy of both the Caprini (P = 0.00051) and PADUA (P = 0.00031) for mortality. ROC for the three- and six-month follow-ups suggested that the Caprini model exhibited superior predictive efficacy for mortality. Conclusions: The PADUA model can serve as a simple and effective tool for stratifying high-risk gynecological inpatients before undergoing CTPA. The Caprini model demonstrated superior predictive efficacy for mortality rates. [ABSTRACT FROM AUTHOR]

COVID-19 prediction has been essential in the aid of prevention and control of the disease. The motivation of this case study is to develop predictive models for COVID-19 cases and deaths based on a cross-sectional data set with a total of 28,955 observations and 18 variables, which is compiled from 5 data sources from Kaggle. A two-part modeling framework, in which the first part is a logistic classifier and the second part includes machine learning or statistical smoothing methods, is introduced to model the highly skewed distribution of COVID-19 cases and deaths. We also aim to understand what factors are most relevant to COVID-19's occurrence and fatality. Evaluation criteria such as root mean squared error (RMSE) and mean absolute error (MAE) are used. We find that the two-part XGBoost model perform best with predicting the entire distribution of COVID-19 cases and deaths. The most important factors relevant to either COVID-19 cases or deaths include population and the rate of primary care physicians. [ABSTRACT FROM AUTHOR]

During the restoration of iron cultural relics, the removal of rust from these artifacts is necessary. However, this rust removal process may lead to inconsistent local color on the iron relics. To address this, mending materials are applied to treat the surface, ensuring consistent local color. In the surface treatment of iron cultural relics, a significant challenge lies in modulating the color of these mending materials. The corrosion products of Yuquan Iron Pagoda are mainly Fe3O4, γ-FeO(OH), α-FeO(OH) and α-Fe2O3, with contents of 13.1, 16.1, 40.2 and 30.6%, respectively. Due to their structural stability and suitable color characteristics, Fe3O4 and α-Fe2O3 are selected as the primary raw materials for the repair material. This study employs machine learning methods to predict the color of mending materials corresponding to varying contents of α-Fe2O3, Fe3O4, and epoxy resin. The Artificial Neural Network (ANN), eXtreme Gradient Boosting (XGBoost), and Light Gradient Boost Machine (LightGBM) algorithms are utilized to develop the model, and the predictive performance of these three algorithms is compared. XGBoost exhibits the best prediction performance, achieving a square correlation coefficient (R2) of 0.94238 and a mean absolute error (MAE) of 0.68485. Additionally, the SHapley Additive exPlanations (SHAP) method is employed to analyze the most crucial raw material affecting the color of mending materials, which is identified as Fe3O4. The study illustrates the specific process of employing this model by applying it to the surface treatment of the Yuquan Iron Pagoda, demonstrating the practicality of the model. This model can be applied to assist in the surface treatment of other iron cultural relics. [ABSTRACT FROM AUTHOR]

Background: Allogeneic Blood transfusion is common in hip surgery but is associated with increased morbidity. Accurate prediction of transfusion risk is necessary for minimizing blood product waste and preoperative decision-making. The study aimed to develop machine learning models for predicting perioperative blood transfusion in hip surgery and identify significant risk factors. Methods: Data of patients undergoing hip surgery between January 2013 and October 2021 in the Peking Union Medical College Hospital were collected to train and test predictive models. The primary outcome was perioperative red blood cell (RBC) transfusion within 72 h of surgery. Fourteen machine learning algorithms were established to predict blood transfusion risk incorporating patient demographic characteristics, preoperative laboratory tests, and surgical information. Discrimination, calibration, and decision curve analysis were used to evaluate machine learning models. SHapley Additive exPlanations (SHAP) was performed to interpret models. Results: In this study, 2431 hip surgeries were included. The Ridge Classifier performed the best with an AUC = 0.85 (95% CI, 0.81 to 0.88) and a Brier score = 0.21. Patient-related risk factors included lower preoperative hemoglobin, American Society of Anesthesiologists (ASA) Physical Status > 2, anemia, lower preoperative fibrinogen, and lower preoperative albumin. Surgery-related risk factors included longer operation time, total hip arthroplasty, and autotransfusion. Conclusions: The machine learning model developed in this study achieved high predictive performance using available variables for perioperative blood transfusion in hip surgery. The predictors identified could be helpful for risk stratification, preoperative optimization, and outcomes improvement. [ABSTRACT FROM AUTHOR]

Background: MicroRNAs (miRNAs) emerge in various organisms, ranging from viruses to humans, and play crucial regulatory roles within cells, participating in a variety of biological processes. In numerous prediction methods for miRNA-disease associations, the issue of over-dependence on both similarity measurement data and the association matrix still hasn't been improved. In this paper, a miRNA-Disease association prediction model (called TP-MDA) based on tree path global feature extraction and fully connected artificial neural network (FANN) with multi-head self-attention mechanism is proposed. The TP-MDA model utilizes an association tree structure to represent the data relationships, multi-head self-attention mechanism for extracting feature vectors, and fully connected artificial neural network with 5-fold cross-validation for model training. Results: The experimental results indicate that the TP-MDA model outperforms the other comparative models, AUC is 0.9714. In the case studies of miRNAs associated with colorectal cancer and lung cancer, among the top 15 miRNAs predicted by the model, 12 in colorectal cancer and 15 in lung cancer were validated respectively, the accuracy is as high as 0.9227. Conclusions: The model proposed in this paper can accurately predict the miRNA-disease association, and can serve as a valuable reference for data mining and association prediction in the fields of life sciences, biology, and disease genetics, among others. [ABSTRACT FROM AUTHOR]

Background: Clinical prediction models can help identify high-risk patients and facilitate timely interventions. However, developing such models for rare diseases presents challenges due to the scarcity of affected patients for developing and calibrating models. Methods that pool information from multiple sources can help with these challenges. Methods: We compared three approaches for developing clinical prediction models for population screening based on an example of discriminating a rare form of diabetes (Maturity-Onset Diabetes of the Young - MODY) in insulin-treated patients from the more common Type 1 diabetes (T1D). Two datasets were used: a case-control dataset (278 T1D, 177 MODY) and a population-representative dataset (1418 patients, 96 MODY tested with biomarker testing, 7 MODY positive). To build a population-level prediction model, we compared three methods for recalibrating models developed in case-control data. These were prevalence adjustment ("offset"), shrinkage recalibration in the population-level dataset ("recalibration"), and a refitting of the model to the population-level dataset ("re-estimation"). We then developed a Bayesian hierarchical mixture model combining shrinkage recalibration with additional informative biomarker information only available in the population-representative dataset. We developed a method for dealing with missing biomarker and outcome information using prior information from the literature and other data sources to ensure the clinical validity of predictions for certain biomarker combinations. Results: The offset, re-estimation, and recalibration methods showed good calibration in the population-representative dataset. The offset and recalibration methods displayed the lowest predictive uncertainty due to borrowing information from the fitted case-control model. We demonstrate the potential of a mixture model for incorporating informative biomarkers, which significantly enhanced the model's predictive accuracy, reduced uncertainty, and showed higher stability in all ranges of predictive outcome probabilities. Conclusion: We have compared several approaches that could be used to develop prediction models for rare diseases. Our findings highlight the recalibration mixture model as the optimal strategy if a population-level dataset is available. This approach offers the flexibility to incorporate additional predictors and informed prior probabilities, contributing to enhanced prediction accuracy for rare diseases. It also allows predictions without these additional tests, providing additional information on whether a patient should undergo further biomarker testing before genetic testing. [ABSTRACT FROM AUTHOR]

Identifying active compounds for target proteins is fundamental in early drug discovery. Recently, data-driven computational methods have demonstrated promising potential in predicting compound activities. However, there lacks a well-designed benchmark to comprehensively evaluate these methods from a practical perspective. To fill this gap, we propose a Compound Activity benchmark for Real-world Applications (CARA). Through carefully distinguishing assay types, designing train-test splitting schemes and selecting evaluation metrics, CARA can consider the biased distribution of current real-world compound activity data and avoid overestimation of model performances. We observed that although current models can make successful predictions for certain proportions of assays, their performances varied across different assays. In addition, evaluation of several few-shot training strategies demonstrated different performances related to task types. Overall, we provide a high-quality dataset for developing and evaluating compound activity prediction models, and the analyses in this work may inspire better applications of data-driven models in drug discovery. Data-driven computational methods have demonstrated promising potential in predicting compound activities from chemical structures, however, unbiased practical applications remain challenging due to the lack of proper benchmarking methods. Here, the authors develop a benchmark termed CARA to eliminate the biases in current compound activity data by using new train-test splitting schemes and evaluation metrics, revealing accurate and informative model performances. [ABSTRACT FROM AUTHOR]

Identifying interactions between T-cell receptors (TCRs) and immunogenic peptides holds profound implications across diverse research domains and clinical scenarios. Unsupervised clustering models (UCMs) cannot predict peptide-TCR binding directly, while supervised predictive models (SPMs) often face challenges in identifying antigens previously unencountered by the immune system or possessing limited TCR binding repertoires. Therefore, we propose HeteroTCR, an SPM based on Heterogeneous Graph Neural Network (GNN), to accurately predict peptide-TCR binding probabilities. HeteroTCR captures within-type (TCR-TCR or peptide-peptide) similarity information and between-type (peptide-TCR) interaction insights for predictions on unseen peptides and TCRs, surpassing limitations of existing SPMs. Our evaluation shows HeteroTCR outperforms state-of-the-art models on independent datasets. Ablation studies and visual interpretation underscore the Heterogeneous GNN module's critical role in enhancing HeteroTCR's performance by capturing pivotal binding process features. We further demonstrate the robustness and reliability of HeteroTCR through validation using single-cell datasets, aligning with the expectation that pMHC-TCR complexes with higher predicted binding probabilities correspond to increased binding fractions. HeteroTCR extracts information on within-type (TCR-TCR or peptide-peptide) similarity and between-type (peptide-TCR) interaction from peptides and TCR input. It is robust across diverse datasets, affirming its potential in immunological applications. [ABSTRACT FROM AUTHOR]

Understanding the dynamics of intracellular signaling pathways, such as ERK1/2 (ERK) and Akt1/2 (Akt), in the context of cell fate decisions is important for advancing our knowledge of cellular processes and diseases, particularly cancer. While previous studies have established associations between ERK and Akt activities and proliferative cell fate, the heterogeneity of single-cell responses adds complexity to this understanding. This study employed a data-driven approach to address this challenge, developing machine learning models trained on a dataset of growth factor-induced ERK and Akt activity time courses in single cells, to predict cell division events. The most predictive models were developed by applying discrete wavelet transforms (DWTs) to extract low-frequency features from the time courses, followed by using Ensemble Integration, a data integration and predictive modeling framework. The results demonstrated that these models effectively predicted cell division events in MCF10A cells (F-measure=0.524, AUC=0.726). ERK dynamics were found to be more predictive than Akt, but the combination of both measurements further enhanced predictive performance. The ERK model's performance also generalized to predicting division events in RPE cells, indicating the potential applicability of these models and our data-driven methodology for predicting cell division across different biological contexts. Interpretation of these models suggested that ERK dynamics throughout the cell cycle, rather than immediately after growth factor stimulation, were associated with the likelihood of cell division. Overall, this work contributes insights into the predictive power of intra-cellular signaling dynamics for cell fate decisions, and highlights the potential of machine learning approaches in unraveling complex cellular behaviors. [ABSTRACT FROM AUTHOR]

Objective Methods Results Conclusion This study aimed to develop and validate a prenatal nomogram to predict the risk of placenta accreta spectrum (PAS) in women with one previous cesarean delivery.This retrospective study enrolled 5157 pregnant women with one previous cesarean delivery in China from January 2021 to January 2023. The nomogram was developed from a training cohort of 3612 pregnant women and tested on a validation cohort of 1545 pregnant women. Multivariate regression analysis was performed using the minimum value of the Akaike information criterion to select prognostic factors that can be included in the nomogram. We evaluated the nomogram by the area under the receiver operating characteristic (ROC) curve, calibration curves, and the decision curve analysis (DCA).PAS occurred in 199 (5.51%) and 80 (5.18%) patients in the training and validation cohorts, respectively. Backward stepwise algorithms in the multivariable logistic regression model determined abortion, hypertensive disorders complicating pregnancy, fetal position, and placenta previa as relevant PAS predictors. The area under the ROC curve for the nomogram was 0.770 (95% confidence interval [CI] 0.733–0.807) and 0.791 (95% CI 0.730–0.853) for the training and validation cohorts, respectively. The calibration curves indicated that the nomogram's prediction probability was consistent with the actual probability. The DCA curve revealed that the nomogram has potential clinical benefit.A prenatal nomogram was developed for PAS in our study, which helped obstetricians determine potential patients with PAS and make sufficient preoperative preparation to reduce maternal and neonatal complications. [ABSTRACT FROM AUTHOR]

Given the ever-evolving role of school counselors and increasing demands placed on them in today's sociopolitical climate, proficiency in program evaluation is more important than ever as school counselors advocate for their role and for comprehensive school counseling program delivery in schools. This article presents a reexamination of the purpose and importance of program evaluation and the development of an evaluator identity. This quantitative study investigated which factors contribute to school counselors' identity as a program evaluator using self-report survey data on levels of self-efficacy, confidence, perceived importance of program evaluation, and interest in program evaluation. It also assessed the overall significance of each factor for engaging in program evaluation behaviors, as operationalized through data-based decision-making practices. Results include a predictive model that displays the contribution and significance of these factors on school counselor program evaluation and related data-based decision-making behaviors. [ABSTRACT FROM AUTHOR]

Mindfulness‐based stress reduction (MBSR) is a 9‐session group‐treatment programme for managing stress. Research suggests variability in the outcomes of MBSR among participants. This prognostic (not causal) study develops a multivariable model that may support clinicians in forecasting expected MBSR outcomes. We used data of 763 patients collected from MBSR programs conducted between October 2015 and March 2022. Candidate prognostic factors at baseline included psychosocial work environment, sociodemographic, and clinical information. Multiple imputation was used to handle missing data (imputations = 200). Important prognostic factors were backward selected in ≥5% of the imputed datasets. The final prediction model including the selected prognostic factors was evaluated using linear regression with a four‐fold internal cross‐validation procedure. Reductions in perceived stress from baseline to end of the MBSR programme were predicted by a lower General Severity Index (β = 2.00, p < 0.01), higher baseline levels of stress (β = −0.88, p < 0.01), and somewhat by having managerial responsibility in the latest job (vs. no; β = −2.53, p = 0.07). The remaining prognostic factors were weaker predictors, for example, gender and income. Internal validity of the final model was indicated by consistent results from four randomly folded subsamples. This study developed a prognostic model predicting changes in stress levels in relation to the MBSR programme. A reduction in stress level was particularly predicted by milder psychological symptoms and higher baseline levels of perceived stress. These predictions cannot be taken as evidence of causal associations. Forecasting of the illness course should be cautiously practiced using clinical judgement regarding individual patients. [ABSTRACT FROM AUTHOR]

Background: Delayed extubation and transfer to the intensive care unit (ICU) in children undergoing major scoliosis surgery may increase postoperative complications, prolong hospital stay, and increase medical expenses; however, whether a child will require delayed extubation or transfer to the ICU after scoliosis orthopedic surgery is not fully understood. In this study, we reviewed the risk factors for delayed extubation and transfer to the ICU after scoliosis orthopedic surgery in children. Method: The electronic medical records of pediatric patients (≤ 18 years) who underwent posterior spinal fusion surgery between January 2018 and November 2021 were reviewed and analyzed. Patient characteristics (age, sex, body mass index, American Society of Anesthesiologists, ASA, grade, preoperative lung function, and congenital heart disease), preoperative Cobb angle, scoliosis type, correction rate, vertebral fusion segments, pedicle screws, surgical osteotomy, intraoperative bleeding, intraoperative allogeneic transfusion, intraoperative hemoglobin changes, intraoperative mean arterial pressure changes, intraoperative tidal volume (ml/kg predicted body weight), surgical time, postoperative extubation, and transfer to the ICU were collected. The primary outcomes were delayed extubation and transfer to the ICU. Multivariate logistic regression models were used to determine the risk factors for delayed extubation and ICU transfer. Results: A total of 246 children who satisfied the inclusion criteria were enrolled in this study, of whom 23 (9.3%) had delayed extubation and 81 (32.9%) were transferred to the ICU after surgery. High ASA grade (odds ratio [OR] 5.42; 95% confidence interval [CI] 1.49–19.78; p = 0.010), high Cobb angle (OR 1.04; 95% CI 1.02–1.07; p < 0.001), moderate to severe pulmonary dysfunction (OR 10.9; 95% CI 2.00–59.08; p = 0.006) and prolonged surgical time (OR 1.01; 95% CI 1.00–1.03; p = 0.040) were risk factors for delayed extubation. A high Cobb angle (OR 1.02; 95% CI 1.01–1.04; p = 0.004), high intraoperative bleeding volume (OR 1.06; 95% CI 1.03–1.10; p = 0.001), allogeneic transfusion (OR 3.30; 95% CI 1.24–8.83; p = 0.017) and neuromuscular scoliosis (OR 5.38; 95% CI 1.59–18.25; p = 0.007) were risk factors for transfer to the ICU. A high Cobb angle was a risk factor for both delayed extubation and ICU transfer. Age, sex, body mass index, number of vertebral fusion segments, correction rate, and intraoperative tidal volume were not associated with delayed postoperative extubation and ICU transfer. Conclusion: The most common risk factor for delayed extubation and ICU transfer in pediatric patients who underwent posterior spinal fusion was a high Cobb angle. Determining risk factors for a poor prognosis may help optimize perioperative respiratory management strategies and planning of postoperative care for children undergoing complicated spinal surgery. [ABSTRACT FROM AUTHOR]

Drug‐drug interaction (DDI) event prediction is a challenging problem, and accurate prediction of DDI events is critical to patient health and new drug development. Recently, many machine learning‐based techniques have been proposed for predicting DDI events. However, most of the existing methods do not effectively integrate the multidimensional features of drugs and provide poor mitigation of noise to get effective feature information. To address these limitations, we propose a DDI‐Transform neural network framework for DDI event prediction. In DDI‐Transform, we design a drug structure information feature extraction module and a drug bind‐protein feature extraction module to obtain multidimensional feature information. A stack of DDI‐Transform layers in the DDI‐Transform network module are then used for adaptive learning, thus adaptively selecting the effective feature information for prediction. The results show that DDI‐Transform can accurately predict DDI events and outperform the state‐of‐the‐art models. Results on different scale datasets confirm the robustness of the method. [ABSTRACT FROM AUTHOR]

Effective clinical trials are necessary for understanding medical advances but early termination of trials can result in unnecessary waste of resources. Survival models can be used to predict survival probabilities in such trials. However, survival data from clinical trials are sparse, and DeepSurv cannot accurately capture their effective features, making the models weak in generalization and decreasing their prediction accuracy. In this paper, we propose a survival prediction model for clinical trial completion based on the combination of denoising autoencoder (DAE) and DeepSurv models. The DAE is used to obtain a robust representation of features by breaking the loop of raw features after autoencoder training, and then the robust features are provided to DeepSurv as input for training. The clinical trial dataset for training the model was obtained from the ClinicalTrials.gov dataset. A study of clinical trial completion in pregnant women was conducted in response to the fact that many current clinical trials exclude pregnant women. The experimental results showed that the denoising autoencoder and deep survival regression (DAE‐DSR) model was able to extract meaningful and robust features for survival analysis; the C‐index of the training and test datasets were 0.74 and 0.75 respectively. Compared with the Cox proportional hazards model and DeepSurv model, the survival analysis curves obtained by using DAE‐DSR model had more prominent features, and the model was more robust and performed better in actual prediction. [ABSTRACT FROM AUTHOR]