Prediction Performance Of Model Research Articles

Stroke-SCORE: Personalizing Acute Ischemic Stroke Treatment to Improve Patient Outcomes.

Background/Objectives: Acute ischemic stroke (AIS) is a leading cause of disability and mortality worldwide. Despite advances in interventions such as thrombolysis (TL) and mechanical thrombectomy (MT), current treatment protocols remain largely standardized, focusing on general eligibility rather than individual patient characteristics. To address this gap, we introduce the Stroke-SCORE (Simplified Clinical Outcome Risk Evaluation), a predictive tool designed to personalize AIS management by providing data-driven, individualized recommendations to optimize treatment strategies and improve patient outcomes. Methods: The Stroke-SCORE was derived using retrospective data from 793 AIS patients admitted to the University of Pécs (February 2023-September 2024). Logistic regression analysis identified age, National Institutes of Health Stroke Scale (NIHSS) score at admission, and pre-morbid modified Rankin Scale (pre-mRS) score as key predictors of unfavorable outcomes at 90 days (defined as modified Rankin Scale [mRS] score > 2). Based on these predictors, a simplified risk score was developed to stratify patients into low-, moderate-, and high-risk groups, guiding treatment decisions on TL, MT, combination therapy (TL + MT), or standard care (SC). Internal validation was performed to assess the model's predictive performance via receiver operating characteristic (ROC) analysis and isotonic regression calibration with bootstrapping. Results: The Stroke-SCORE was moderately positively correlated with a 90-day mRS score > 2 (odds ratio [OR] = 0.70, 95% confidence interval [CI]: 0.58-0.83, p < 0.001), with an area under the curve (AUC) of 0.86, a sensitivity and specificity of 79% and 81%, respectively, and an overall accuracy of 80%. Simulations indicated that personalized treatment guided by the Stroke-SCORE significantly reduced unfavorable outcomes. Conclusions: The Stroke-SCORE demonstrates strong predictive performance as a practical, data-driven approach for personalizing AIS treatment decisions. In the future, external, multicenter prospective validation is needed to confirm its applicability in real-world settings.

Predicting intraocular pressure using neural networks: Incorporating image data and patient metadata from papila dataset

Aims/Purpose: This study aimed to explore the feasibility of predicting intraocular pressure (IOP) using a neural network model incorporating both image data and patient metadata. The objective was to evaluate the potential of this approach in improving the understanding and prediction of IOP levels in individuals, utilizing data from the PAPILA public database.Methods: We employed a dataset containing 331 images from the PAPILA database, with 298 images classified as normal and 33 diagnosed with glaucoma. Additionally, patient metadata such as glaucoma diagnosis status, age, sex, refractive defect, pachymetry, and axial length were incorporated. The dataset was divided into 70% for training the neural network model, 15% for validation, and 15% for testing.Results: The analysis of the neural network model's predictive performance produced the following metrics: Mean Absolute Error (MAE) of 2.515, Root Mean Square Error (RMSE) of 3.131, and Mean Absolute Percentage Error (MAPE) of 16.611%. These results indicate a moderate level of predictive capability for IOP based on the input data.Conclusions: In conclusion, our study demonstrates promising potential in utilizing artificial intelligence to predict intraocular pressure (IOP) using image data and patient metadata. While our model's predictive performance is moderate, it sets the stage for further advancements. With continued refinement and exploration of additional factors, AI‐driven approaches hold promise for revolutionizing glaucoma management.References Kovalyk O, Morales‐Sánchez J, Verdú‐Monedero R, Sellés‐Navarro I, Palazón‐Cabanes A, Sancho‐Gómez JL. PAPILA: Dataset with fundus images and clinical data of both eyes of the same patient for glaucoma assessment. Sci Data. 2022 Jun 9;9(1):291. doi: 10.1038/s41597‐022‐01388‐1. PMID: 35680965; PMCID: PMC9184612.

PSO/GA/ANN modeling and prediction for the higher heating values (HHVS) of solid fuels: The machine learning approach

The quick evaluation for the higher heating value (HHV) is crucial for thermochemical conversion of solid fuels. In this work, machine learning method based on artificial neural networks (ANN) was used to predict the HHV of solid fuel. 205 groups of different kinds of solid fuels collected from publications were used. The proximate analysis, ultimate analysis and the combination of two were used as input parameters. The influence of activation function, neuron number and hidden layer number on the prediction performance was studied. Results show that single hidden layer with logsig function using 8 neurons was an optimized condition for HHV prediction. The combination of two composition analyses could achieve much higher accuracy, with the average relative error of 2.57%. Impact analysis indicated that the non-combustible components, namely ash content and oxygen content showed the largest influencing weight for HHV prediction, accounting for 21.73% and 22.91% respectively. Particle swarm optimization (PSO) and genetic algorithm (GA) were further used to optimize the artificial neural network model. Results show that PSO and GA both improved the prediction performance of ANN model by optimizing the initial weight and threshold values. The average relative errors for PSOANN and GA-ANN decreased to 1.15 % and 1.72 % respectively.

Prediction of load-bearing capacity of FRP-steel composite tubed concrete columns: using explainable machine learning model with limited data

The complex interaction mechanism between the FRP jacket, steel tube, and confined concrete in the FRP-steel composite tubed concrete (F-STC) column makes the prediction of its mechanical behaviour a challenging task. This paper specifically investigates the application of machine learning models to predict the load-bearing capacity of F-STC columns. Since relatively few experimental works are performed on F-STC columns, the database established based on the test results from previous research contains only 69 data samples. In order to circumvent the training difficulties caused by the relatively small database, the application of the Gaussian process regression (GPR) model is trialled in this study. To make the predictions of the GPR model explainable, the Shapley additive explanations (SHAP) approach is incorporated with the GPR model in this study. Besides, four contrasting prediction models based on artificial neural network (ANN), support vector regression (SVR), decision tree (DT), and random forest (RF), are also proposed. The k-fold validation and test results indicate that the GPR model provides strong potential in predicting the load-bearing capacity of F-STC columns with high prediction accuracy and generalisation capability. Besides, 95% and 99% confidence intervals obtained by the GPR model are provided to show the uncertainty of the prediction results. Furthermore, the effect of the database size on the prediction performance of GPR, ANN, and SVR models is further examined by gradually reducing the number of data samples, and the comparisons illustrate the superiority of the GPR model.

CXCL5 as a biomarker for early diagnosis and prognosis of sepsis: A comprehensive clinical evaluation.

Sepsis, a critical condition caused by a dysregulated host response to infection, has high morbidity and mortality rates. Timely diagnosis and treatment are vital for improving patient outcomes. This study explores the potential role of CXCL5 in the diagnosis, severity assessment, and prognosis of sepsis. We included 147 sepsis patients, 50 patients with systemic inflammatory response syndrome (SIRS) and 120 healthy controls. Serum CXCL5 levels, inflammation scores (APACHE II, SOFA), and other laboratory indicators were recorded. Univariate and multivariate logistic regression analyses were conducted to assess the relationship between CXCL5 and sepsis diagnosis, severity, and prognosis. A prognostic nomogram was constructed and evaluated using receiver operator characteristic curves, calibration curves, and clinical decision curves. Serum CXCL5 levels in sepsis patients were significantly higher than those in patients with SIRS and healthy controls. CXCL5 was identified as a risk factor for sepsis diagnosis. CXCL5 levels were significantly elevated in patients with septic shock (P=0.04) and in deceased patients compared to survivors (P<0.001). The prognostic model, incorporating CXCL5, lactate, APACHE II scores, C-reactive protein levels, and respiratory rate, demonstrated high predictive accuracy with an area under the curve of 0.873. Calibration and decision curve analyses demonstrated the model's good predictive performance and potential clinical value. Serum CXCL5 concentration is a promising biomarker for enhancing the diagnostic accuracy and prognostic evaluation of sepsis. The constructed multivariate prediction model offers new insights into sepsis prognosis, but its direct application in clinical practice requires further validation.

Towards visual field reliability with machine learning

Aims/Purpose: To evaluate the reliability of visual field mean deviation slope with explainable artificial intelligenceMethods: 6,869 visual field reports (Humphrey Visual Field, Carl Zeiss) from a study of 645 participants with glaucoma were acquired using a 24‐2 SITA FAST protocol. An additional 704 visual field reports from 54 glaucoma patients with the same visual field protocol was also retrospectively acquired from the clinic. The visual field data was organized into 5‐visit segments and visual field mean deviation (MD) slope was calculated for each segment. 3‐visit MD slopes from corresponding segments aligned with the first visit of each 5‐visit segment were also obtained. A gradient‐boosted machine learning model was developed to predict if the difference between the 3‐visit and 5‐visit slopes was within 1 dB/year. 2,883 segments from the study cohort were used for model development. Area under the receiver operating curve was used to evaluate the model prediction performance, which was compared with a logistic regression baseline operating on the slope values. Model explainability was evaluated using Shapley additive explanations (SHAP).Results: The models were evaluated on 815 segments from the study cohort and 449 from the clinical cohort. Of these, 221 from the study cohort (27%) and 204 (45%) from the clinical cohort were found to have MD slope differences within 1 dB/yr between the 3‐visit and 5‐visit segments. The machine learning model achieved AUCs of 0.89 and 0.88 on the study and clinical cohorts, compared to 0.87 and 0.85 respectively with logistic regression. SHAP analysis found that the key determinants of the model were the magnitudes of the slope values, followed by inter‐visit intervals.Conclusions: The machine learning model better predicted visual field slope reliability compared to logistic regression. MD slope values and inter‐visit intervals were found to be important factors in VF reliability analysis.

Bridge Superstructure Component Performance Prediction Model Based on LightGBM

The rapid increase in the number of aging bridges domestically requires efficient maintenance strategies, which may utilize bridge performance prediction models. In this study, performance prediction models were developed for structural components with large amounts of data in the superstructure (concrete decks, PSC I girders, steel box girders, concrete crossbeams, and steel crossbeams) using data managed by a Bridge Management System. The condition rating from previous inspections (previous rating) had the greatest impact on the predictions, and regression analysis was used to train the model with the regression value of the previous rating rather than its representative value. Consequently, the prediction performance of the learning model improved, and the same method was applied to other superstructure components. Finally, the prediction results were analyzed by applying actual bridge data to the model.

Marketing and Project Management 4.0: Next-Generation Digital Tools

Modern marketing and project management are now largely dependent on artificial intelligence (AI), which helps companies make better decisions, streamline operations, and increase profitability. This study examines sophisticated mathematical models to assess AI's performance in several domains. AI model accuracy, predictive performance, and impact are evaluated using key performance measures such as Mean Squared Error (MSE), Log-Loss, and Improvement over Baseline (IOB). The study also presents AI Training Cost Functions (CTF) to examine model training expenses and Expected Profit (EP) to quantify the monetary gains of AI-driven campaigns. These algorithms allow us to measure how AI contributes to increased productivity, higher conversion rates, and more efficient use of resources. The study also looks at how, in comparison to conventional approaches, AI-powered marketing campaigns increase revenue, customer retention, and engagement. This article shows how companies may use AI to improve project management workflows and maximize marketing efforts, making them more successful and economical, using real-world examples. The results indicate that AI plays a crucial role in project management and marketing, offering quantifiable financial gains as well as operational enhancements.

Development and validation of a novel model to predict post-stroke cognitive impairment within 6 months after acute ischemic stroke.

Cognitive decline following acute ischemic stroke (AIS), termed post-stroke cognitive impairment (PSCI), is a prevalent phenomenon that significantly elevates disability and mortality rates among affected patients. The objective of this investigation was to develop a robust clinical prediction model capable of forecasting PSCI within six months post-AIS and subsequently validate its effectiveness. A cohort of 573 AIS patients was stratified into two groups: those with PSCI (260 cases) and those who remained cognitively normal (CN) (313 cases). These patients were further subdivided into three distinct cohorts: a development cohort comprising 193 AIS patients, an internal validation cohort with 193 AIS patients, and an external validation cohort encompassing 187 AIS patients. A thorough multifactor logistic regression analysis was conducted to identify independent predictors of PSCI, which were subsequently incorporated into the prediction model for comprehensive analysis and validation. The discriminatory power, calibration accuracy, and clinical net benefits of the prediction model were rigorously evaluated using the area under the receiver operating characteristic curve (AUC-ROC), calibration plots, and decision curve analyses, respectively. Utilizing a meticulously selected panel of variables, including smoking status, alcohol consumption, female gender, low educational attainment, NIHSS score at admission, stroke progression, diabetes mellitus, atrial fibrillation, stroke localization, HCY levels, and Lp-PLA2 levels, a clinical prediction model was formulated to predict the occurrence of PSCI within six months of AIS. The model demonstrated AUC-ROC values of 0.898 (95%CI, 0.853-0.942), 0.847 (95%CI, 0.794-0.901), and 0.849 (95%CI, 0.7946-0.9031) in the development, internal validation, and external validation cohorts, respectively. Further validation through calibration curve analyses, Hosmer-Lemeshow goodness-of-fit tests, and additional metrics confirmed the model's impressive predictive performance. The proposed model exhibits strong discriminative ability for predicting PSCI and holds considerable promise for guiding clinical decision-making. However, ongoing optimization with multicenter data is necessary to bolster its robustness and broaden its applicability.

The risk factors and prediction model for postoperative pneumonia after craniotomy.

Craniotomy is highly susceptible to postoperative pneumonia, which significantly impacts the outcomes of patients undergoing such procedures. Our study aims to examine the risk factors associated with postoperative pneumonia and establish a predictive model with a nomogram to assess this risk. We conducted a matched 1:1 case-control study involving 831 adult patients undergoing craniotomy at our hospital. Cases consisted of patients who developed postoperative pneumonia within 30 days after surgery, as defined by consensus criteria. Controls were randomly selected from a pool of eligible patients. The overall incidence rate of postoperative pneumonia is 12.39% in a total of 831 surgeries, which associated with unfavorable outcomes. Gram-negative bacteria were found to be the most common causative agents and approximately 27.94% of cases attributed to multi-drug resistant strains. The logistic regression analysis revealed five independent risk factors, as follows: smoking history, surgical duration, postoperative albumin, unplanned re-operation, and deep vein catheterization. A risk prediction model was derived and a nomogram was constructed. The Hosmer-Lemeshow test yielded X2 = 3.871 (P=0.869), and the receiver operator characteristic curve analysis demonstrated an area under the curve of 0.898 (P<0.05), with a sensitivity of 79.6% and a specificity of 85.4%, indicating excellent model fit and predictive performance. In addition, the C-index of the nomogram model was 0.898(95%CI, 0.853~0.941). The calibration curves of the nomogram model showed p-values of 0.797 and the Brier scores were 0.127. The analysis of the clinical decision curve showed that the nomograph model had high clinical application value. Postoperative pneumonia patients after craniotomy exhibits distinct pathogen distribution and is strongly associated with unfavorable outcomes. The risk prediction model developed in this study demonstrates a good fitting degree and predictive performance. The constructed nomogram model is objective, specific, and easily applicable in clinical practice.

Balancing model specificity and transferability: Barn owl nest box selection

AbstractExamining the transferability of habitat selection models is vital when they are used to forecast predictions in new times or places, but this issue is too often neglected. Nest boxes are often installed in agricultural landscapes to attract barn owls (Tyto spp.) and the ecosystem services they provide. For this practice to be effective, farmers need actionable guidelines for nest box design and placement to optimize nest box use. We addressed 3 primary objectives: 1) develop a nest box selection model in the well‐studied region of Napa Valley, California, USA, 2) evaluate this model's predictive performance in other regions of California, and 3) use data from all regions to build a more generalizable model. Based on data from 6 years of monitoring used and unused American barn owl (Tyto furcata) nest boxes in Napa Valley, we found that nest box selection was best predicted by nest box attributes (e.g., pole height, box height, and entrance orientation), local land cover (e.g., grassland within 75 m), and landscape‐scale metrics (e.g., grassland within 2.81 km). This model's predictions were strongly correlated with observed nest box use in Napa, but the model performed poorly when used to predict nest box use in other regions that are ecologically similar (Sonoma County) or dissimilar (Fresno, Merced, and Madera counties). A model pooling data from all regions fit the data well and again showed effects of box, local, and landscape predictors. It was more generalizable than the Napa‐only model and lost little precision when applied with forecasting predictions to Napa in particular. Taken together, our results indicate that local data should be used to make the most reliable predictions of nest box use. Until those data are available, general recommendations should be made from models that pool data from as many regions as feasible and should provide appropriate caveats. Results of this work can inform nest box design and placement for the benefit of farmers and owls in California, and future research should examine nest box selection by barn owls in other areas of the world with different climates and local habitats.

가정환경 및 인구학적 요인에 따른 11세 아동의 문제행동 예측모델 개발

Objectives The purpose of this study is to evaluate the performance of a prediction model to determine whether the home environment and demographic factors of 10-year-old children predict problem behaviors one year later and to identify important predictive factors. Methods The research data used the 11th (data collection in 2018) and 12th (data collection in 2019) data of the Korean Children's Panel. The input variables are demographic variables, parenting behavior and mental health-related variables reported by parents, and family relationship variables reported by children in the 11th data of 1,177 10-year-old children (574 girls and 603 boys). The output variable is whether the children at the age of 11 in the 12th data have clinical-level internalizing and externalizing behavior problems (T≥64). The XGBoost algorithm was used to develop the prediction model. Results The analysis results are as follows. First, the prediction performance of the XGBoost models predicting clinical-level problem behaviors was very good, with AUC = 1.00 for both internalizing and externalizing behavior problems. Second, the key predictor of clinical-level problem behaviors was the mother’s perceived marital conflict for internalizing behavior problems, and the mother’s permissive parenting style for externalizing behavior problems. Conclusions The problem behavior prediction model for school-age children, developed using machine learning and data from the Korean Children's Panel, is highly accurate in predicting whether problem behavior will reach clinical levels a year in advance. This highlights the importance of supporting marital relationships and positive parenting practices to prevent problematic behaviors in children.

Predicting lane-changing risk considering class imbalance problem: a control method for synthetic sample

Abstract The class imbalance has long been a problem in traffic safety research, which significantly reduces the prediction accuracy for minority class samples. In this study, a Control Methods for Synthetic Sample (CMSS) is proposed to control samples synthesized in direction which is conducive to the improvement of prediction performance. This method first employs several oversampling methods to synthesize samples. Then, a regression model is established based on the cluster centers of synthetic samples and the corresponding value of performance improvement after adding the synthetic samples. The Shapley Additive Explanations (SHAP) technique is used to interpret it, and thus determine the features that have the most significant influence on the improvement. After that, random disturbance terms are added to several important features to make a directional adjustment. To achieve better prediction performance, the particle swarm algorithm (PSO) is utilized to search for the number of features and the value of random perturbation terms. The highD dataset is used to validate the effectiveness of CMSS by predicting the lane-changing risk. In addition, the impact of factors such as imbalance rates on the effect of CMSS is discussed. Results show that CMSS can effectively improve the prediction performance of the non-integrated learning models under different class imbalance ratios. It is also revealed that the category of the regression model has a significant effect on the results of CMSS. As a novel method, CMSS can be combined with various kinds of oversampling methods to bring new perspectives for solving the class imbalance problem.

Automation of Local Regression Model Building for Spectroscopic Data

ABSTRACTThe concept of using local calibration for spectroscopic analysis has been discussed since the late 1980s. Since that time, many papers have described modifications to different aspects of the local modeling methodology. In this paper, we briefly discuss some of the modifications and describe an approach for the unattended automation of local model development. Ways to reduce calculation time are discussed. Four example spectroscopic datasets using Raman, FT‐NIR, and dispersive NIR are analyzed, and the local model prediction performance is compared to standard PLS prediction performance. Using independent prediction sets, local modeling is shown to improve prediction performance by 17% to 55%.

Machine Learning-based Dissolved Oxygen Prediction Modeling and Evaluation in the Yangtze River Estuary

Dissolved oxygen （DO） serves as a pivotal indicator, mirroring the intrinsic self-purification capacity of aquatic ecosystems and the overarching quality of the water environment. In the context of the Yangtze Estuary, a crucial hub for biodiversity and economic activities in China, understanding and forecasting levels of DO is instrumental for effective environmental stewardship and management strategies. Considering this, the introduction of sophisticated machine learning algorithms into the monitoring and predictive analytics of dissolved oxygen levels represents an important stride toward leveraging the power of data-driven insights for environmental sustainability. The Yangtze Estuary, characterized by its dynamic and complex hydrological and ecological systems, demands an insightful and nuanced approach to monitoring water quality parameters. To this end, six key monitoring stations were chosen across the estuary, including Xuliujing, Nantong Port, Qidong Port, Qinglong Port, South Port, and North Port, acting as sentinel sites for gauging the health of the water body. Leveraging three cutting-edge modeling techniques-particle swarm optimization-support vector regression （PSO-SVR）, artificial neural network （ANN）, and random forest （RF）-the research unraveled and forecasted the patterns of dissolved oxygen levels using monthly average water quality data spanning from 2004 to 2020. These models embodied the forefront of machine learning technology, each bringing distinct analytical strengths and perspectives to the table, from the nuanced, non-linear pattern recognition capabilities of ANN to the robustness and interpretability of RF. The meticulous evaluation conducted via the RF model underscored the paramount importance of three water quality variables, namely temperature, five-day biochemical oxygen demand, and ammonia nitrogen, in influencing the spatial-temporal dynamics of dissolved oxygen in the estuary. Comparative analysis of the prediction results yielded by the PSO-SVR, ANN, and RF models illuminated the superior performance of the RF model across the six monitoring stations, with an overall average error margin of 0.19, a testament to its efficacy and reliability. In comparison, the PSO-SVR and ANN models exhibited higher error rates of 0.38 and 0.47, respectively, albeit still contributing valuable insights into the complex dissolved oxygen dynamics in the Yangtze Estuary. The prediction performance of the machine learning models was evaluated, and the overall prediction performance ranking on the training set was RF （R2=0.971； RMSE=0.341 mg·L-1） &gt; PSO-SVR （R2=0.884； RMSE=0.707 mg·L-1） &gt; ANN （R2=0.792； RMSE=0.967 mg·L-1）. The overall prediction performance ranking on the test set was RF （R2 = 0.986； RMSE=0.165 mg·L-1） &gt; PSO-SVR （R2=0.951； RMSE=0.332 mg·L-1） &gt; ANN （R2=0.800； RMSE=0.633 mg·L-1）. Therefore, the RF model exhibited the best predictive ability on all monitoring sections, showing excellent performance and generalization ability both on the training and the test sets. The PSO-SVR model also performed well on most monitored profiles, with slightly lower predictive performance than that of the RF model though with better stability and generalization ability. However, the ANN model did not perform as perfectly as the other two models in some monitoring profiles and its network structure or parameters may need to be further optimized to improve the prediction accuracy and stability.

Psychological distress in biliary tract malignancy patients: influencing factors and development of a predictive nomogram model.

This study aims to investigate the psychological distress and its influencing factors in patients with biliary tract malignant tumors, alongside the development of a predictive model. A total of 219 patients diagnosed with biliary tract malignant tumors who were admitted to the Department of Liver Surgery at Fudan University Shanghai Cancer Center from July 2021 to May 2023, were selected using a convenience sampling method. Research tools involve psychological distress management screening tools, a demographic questionnaire, self-rating anxiety and depression scales, and the Chinese version of the Memorial Symptom Assessment Scale. Bootstrap method was utilized for repeated sampling to identify relevant factors influencing psychological distress in biliary tract cancer patients. The R software was employed to create a nomogram model, and the model's accuracy and predictive performance were assessed using the receiver operating characteristic curve (ROC) and the Hosmer-Lemeshow test. The average score of psychological distress among the 219 patients was (3.91 ± 2.44), with a psychological distress detection rate of 54.8%. Regression model results indicated that factors such as the presence of distant metastasis, comorbidity with other major diseases, poor sleep quality, anxiety, and severity of anxiety and depression were the primary influencers of psychological distress. The detection rate of psychological distress in patients with biliary tract malignant tumors is notably high. The predictive model constructed in this study exhibits good predictive efficacy and clinical value, providing valuable reference for healthcare professionals in developing targeted intervention strategies.

Repurposing Drugs for Infectious Diseases by Graph Convolutional Network with Sensitivity-Based Graph Reduction.

Computational systems biology employs computational algorithms and integrates diverse data sources, such as gene expression profiles, molecular interactions, and network modeling, to identify promising drug candidates through repurposing existing compounds in response to urgent healthcare needs. This study tackles the urgent need for rapid therapeutic development against emerging infectious diseases. We introduce a novel analytic expression for sensitivity analysis based on the Kronecker product and enhance model prediction performance using Graph Convolutional Networks (GCNs) with sensitivity-based graph reduction. Our algorithm refines prediction performance by leveraging sensitivity-based graph reduction. By integrating RNA-seq data, molecular interactions, and GCNs, we identify disease-related genes and pathways, construct heterogeneous graph models, and predict potential drugs. This approach involves novel analytical expressions that assess sensitivity to model loss, employing the Kronecker product approach. Subgraph analysis identifies nodes for removal, leading to a refined graph used for model retraining. This cost-effective pipeline focuses on computational methods for drug repurposing, targeting infectious diseases such as Zika virus and COVID-19 infection. Applied to these infections, our methodology integrates 659 proteins and 703 drugs for Zika virus, and 495 proteins and 468 drugs for COVID-19, along with their interactions derived from gene expression profiles. Top candidate drugs, such as Betamethasone phosphate and Bizelesin for Zika virus, and Chloroquine, Heparin Disaccharide, and Resveratrol for COVID-19, were validated through literature review or docking analysis. This scalable approach demonstrates promise in repurposing drugs for urgent healthcare challenges.

Radiomic signatures of brain metastases on MRI: utility in predicting pathological subtypes of lung cancer.

The pathological sub-classification of lung cancer is crucial in diagnosis, treatment and prognosis for patients. Quick and timely identification of pathological subtypes from imaging examinations rather than histological tests could help guiding therapeutic strategies. The aim of the study is to construct a non-invasive radiomics-based model for predicting the subtypes of lung cancer on brain metastases (BMs) from multiple magnetic resonance imaging (MRI) sequences. One hundred and sixty-one patients of primary lung cancer with synchronous BMs [121 with adenocarcinoma (AD); 40 with small cell lung carcinoma (SCLC)] were enrolled in the study (129 and 32 in the training set and validation set). A total of 960 radiomics features were extracted from multiple MRI sequences [fluid attenuated inversion recovery (FLAIR), diffusion weighted imaging (DWI), contrast-enhanced T1 weighted imaging (CE-T1WI) and contrast-enhanced susceptibility weighted imaging (CE-SWI)] and four clinical features were recorded. Forty-one key features were selected by the least absolute shrinkage selection operator (LASSO). The machine learning (ML) models for predicting AD and SCLC with radiomics features alone and with radiomics features plus clinical features were constructed using classifiers of logistic regression (LR), random forest (RF), support vector machine (SVM) and extreme gradient boosting (XGBoost). The prediction performance of models was evaluated by accuracy (ACC), sensitivity (SEN), specificity (SPE), F1 score and area under the curves (AUC). The AUCs of LR, RF, SVM and XGBoost models were 0.8177 vs. 0.7604, 0.8177 vs. 0.7839, 0.4792 vs. 0.8594 and 0.9062 vs. 0.8750, respectively, when using radiomics features alone and radiomics features plus clinical features. In the best-performing model using XGBoost, combination of conventional MRI sequences and CE-SWI had better sub-classification performance than conventional MRI sequences. Radiomics of BMs from multiple MRI sequences provides high discriminatory performance in predicting AD and SCLC using classifiers of LR, RF and XGBoost, and can serve as a potential useful tool to non-invasively distinguish pathological subtypes of lung cancer.

A novel nephrometry scoring system outperforms RENAL score, PADUA score, and SPARE score for feasibility prediction of nephron-sparing surgery in children with bilateral Wilms tumor.

Several nephrometry scores have been proposed and applied in the adult population. We previously established a novel model to predict the feasibility of nephron-sparing surgery (NSS) in pediatric bilateral Wilms tumor (WT) patients. This study aimed to evaluate whether our model had better predictive performance compared to other scores. Data from 58 patients during the period 2008-2019 were retrospectively reviewed. Nephrometry scores, including RENAL [radius (R), exophytic/endophytic (E), nearness to collecting system/sinus (N), anterior/posterior (A), and location relative to polar lines (L)], PADUA (preoperative aspects and dimensions used for an anatomical classification), and SPARE (Simplified PADUA REnal) scores, were calculated. A senior radiologist and a senior urologist independently reviewed computed tomography (CT) or magnetic resonance scans while blinded to clinical outcomes. A urology resident re-evaluated the imaging scans using our RSR [relation with collecting system (R), size (S), remaining renal parenchyma (R)] model. The areas under the curve (AUCs) were compared. Inter-rater reliability of the RSR model was assessed using kappa statistics and intraclass correlation coefficient (ICC). The AUC was 0.982 for the RSR model, which outperformed the other three scores. Decision curve analysis (DCA) showed the same results. Compared to the RENAL score, the RSR model showed significant reclassification ability [0.573 (95% confidence interval: 0.161, 1.204), P=0.03]. Significant overall discrimination ability was found with RSR compared to the other three scores (P<0.001). The ICC value was 0.871 for S, and the kappa values of the two R indicators were 0.806 and 0.777, respectively. A web-based calculator was constructed. The RSR model's predictive performance outperforms the established adult nephrometry scores while offering overall fair interobserver agreement. A web-based calculator could help to make personalized treatment decisions.

Predicting the depth of rock cutting by abrasive water jet using support vector machine optimized with whale optimization algorithm

The utilization of abrasive water jet (AWJ) has garnered notable attention in subsurface engineering, as well as unconventional natural gas development, geothermal energy extraction, and tunnel excavation. The efficiency of construction operations is contingent upon rock fragmentation, which is controlled by AWJ parameters and rock properties. Currently, the parameter settings for rock fragmentation by AWJ predominantly rely on empirical approaches, and existing prediction models have large errors due to a limited number of training samples. In this study, we propose a combined support vector machine (SVM) and whale optimization algorithm (WOA) model. To test the model's predictive performance for rock-breaking depth, a database consisting of eight input parameters is constructed. These parameters include AWJ pressure, target distance, lateral velocity, abrasive types, mass flow rate, abrasive particle size, rock types, and rock uniaxial compressive strength. Additionally, to demonstrate the superiority of the WOA-SVM model, three other predictive models based on the back propagation (BP) network, SVM, and Random Forest (RF) are established, compared, and evaluated. The results show that the optimized WOA-SVM model is the most accurate in predicting rock cutting depth, achieving a precision rate of 0.972 25 compared to other models (BP: 0.9536; RF: 0.9681; SVM: 0.9687). Furthermore, sensitivity analysis highlights that lateral velocity exhibits the highest impact on the model, followed by jet pressure and the uniaxial compressive strength of rock. This underscores the critical importance of prioritizing the adjustment of lateral velocity, AWJ pressure, and rock properties when engaging in rock-cutting operations.