Individual Models Research Articles

Computationally integrating radiology and pathology image features for predicting treatment benefit and outcome in lung cancer

Lung cancer, the leading cause of cancer-related deaths globally, includes non-small cell lung cancer (NSCLC) (85% of cases) and small cell lung cancer (SCLC) (13–15%). While accurate diagnosis and treatment selection are critical, the absence of reliable predictive or prognostic biomarkers remains a significant challenge. This study explored the combined use of radiomics from CT scans and pathomics from H&E slides in three contexts: (1) predicting disease recurrence in early-stage NSCLC, (2) predicting immunotherapy response in advanced-stage NSCLC, and (3) predicting chemotherapy response in SCLC. The integrated radio-pathomic model significantly outperformed individual models. In early-stage NSCLC (N = 194), it achieved an HR of 8.35 (C-index: 0.71, p = 0.0043). In advanced-stage NSCLC (N = 35), the combined model improved predictive performance (AUC: 0.75, p = 0.042). In SCLC (N = 50), the integrated model showed an AUC of 0.78, surpassing both radiomic and pathomic models. These findings highlight the potential of combining radiomics and pathomics for improved lung cancer risk stratification and treatment prediction.

Optimized Credit Card Fraud Detection Leveraging Ensemble Machine Learning Methods

The increasing number of online financial transactions, particularly those involving credit cards, underscores the urgent need for robust security systems to mitigate financial losses due to fraud. This paper presents a novel machine learning-based approach to credit card fraud detection that addresses the growing demand for enhanced security. Unlike traditional single-model approaches, the proposed ensemble model uniquely combines random forest, logistic regression, and AdaBoost techniques to accurately distinguish between legitimate and fraudulent transactions. The novelty of this work lies in its innovative soft voting mechanism, which aggregates the strengths of these diverse algorithms to achieve superior classification accuracy and minimize false positives. By leveraging the complementary strengths of each model, the ensemble approach provides a robust and adaptive framework capable of detecting emerging fraud patterns in real-time. The results demonstrate that the proposed ensemble model outperforms individual models, achieving an accuracy of 99.96%, a precision of 99.53%, and a recall of 100%, with an F1 score of 0.99 and an Area Under the Curve (AUC) of 1.0. These findings highlight the model's ability to significantly reduce false positives and negatives, making it a highly reliable solution for fraud detection. Furthermore, the model's performance was validated on the PaySim dataset, a large-scale synthetic dataset, where it achieved a 99.97% accuracy, demonstrating its generalizability and robustness in handling complex, real-world scenarios. This research contributes to the field by introducing a highly extensible and adaptable fraud detection framework that improves current solutions and provides a foundation for future advancements in ensemble learning for fraud detection. The proposed model's ability to integrate multiple classifiers while maintaining interpretability and computational efficiency sets it apart from previous studies, offering a promising direction for enhancing the security of online financial transactions.

High-fidelity multidisciplinary maneuver simulations of the Falcon 2000LX ISTAR

Abstract This paper presents high-fidelity multidisciplinary maneuver simulations of the new research aircraft Dassault Falcon 2000LX ISTAR of the German Aerospace Center (DLR). These simulations are computed using the multidisciplinary process chains FSAerOpt and UltraFLoads for the DLR project HighFly (high-speed inflight validation). The goal of this project is the validation of simulation models for virtual flight testing with data measured on the Falcon 2000LX ISTAR. The paper describes the individual numerical models and analyses of ISTAR: geometry (CAD), aerodynamics (CFD), control surface deflection in CFD, structure (CSM), aircraft mass distribution, and engine. It also gives a short overview of some UltraFLoads simulation processes, such as CFD-CSM coupling, trim analysis, and dynamic maneuver simulation. After that, the simulation results such as steady-state CFD-CSM coupled simulation of different trimmed horizontal flights and simulation of different dynamic maneuvers (bank-to-bank and pull-up and push-down maneuver) are presented. The first comparisons of the numerical simulations with the experimental data are also shown.

Individual Tree Mortality Prediction of Pinus yunnanensis Franch.—Based on Stacking Ensemble Learning and Threshold Optimization

Accurate prediction of individual tree mortality in Pinus yunnanensis Franch. is essential for sustainable forest management and ecological monitoring in southwest China. The aim of this study is to develop a tree mortality prediction model for Pinus yunnanensis based on resurvey data from the Cangshan area in Dali, Yunnan Province, using a stacked ensemble learning algorithm. After an initial evaluation of model performance, the classification thresholds were optimized using the Minimum Classification Error method, the Maximum Sensitivity and Specificity method, the Kappa coefficient method, and the Precision-Recall (PR) curve method to enhance classification results. The findings show that, compared to traditional statistical methods and individual machine learning models, the stacked ensemble learning model (Stacked-RSX) outperforms others in tree mortality classification tasks, which achieved an accuracy of 0.8947, recall of 0.9431, true negative rate of 0.9490, misclassification rate of 0.2289, and an area under the curve of 0.953. Through an exhaustive search for the best classification thresholds, the PR curve method demonstrated good adaptability across all models. All optimal thresholds, relative to the default threshold, significantly improved overall classification performance. Furthermore, feature importance analysis revealed that tree height, diameter at breast height (DBH), Hegyi competition index, and the ratio of DBH to stand basal area are key variables influencing mortality risk. These results indicate that the stacking ensemble learning algorithm effectively analyzes the complex relationships among different factors, significantly improving the prediction accuracy of tree mortality, and providing scientific insights for the management and health monitoring of Pinus yunnanensis forests.

Developing a novel hybrid model based on GRU deep neural network and Whale optimization algorithm for precise forecasting of river’s streamflow

Streamflow contemplates a fundamental criterion to evaluate the impact of human activities and climate changes on the hydrological cycle. In this study, a novel innovative deep neural network (DNN) structure by integrating a double Gated Recurrent Units (GRU) neural network model with a multiplication layer and meta-heuristic whale optimization algorithm (WOA) (i.e., hybrid 2GRU×–WOA model) is developed to improve the prediction accuracy and performance of mean monthly Chehel-Chai River’s streamflow (CCRSFm) in Iran. The Pearson’s correlation coefficient (PCC) and Cosine Amplitude Sensitivity (CAS) as feature (input) selection process determine the only precipitation (Pm) as the most effective input variable among a list of on-site potential climate time series parameters recorded in the study area. Thanks to a well-proportioned layer network structural framework in the suggested hybrid 2GRU×–WOA model, it leads to an appropriate total learnable parameter (TLP) compared to standard individual GRU and Bi-GRU as the benchmark models developed in the comparable meta-parameters. This hybrid model under the optimal meant meta-parameters tuned i.e., coupling a state activation functions (SAF) of tanh-softsign, dropout rate (P-rate) of 0.5, numbers of hidden neurons (NHN) of 70, outperforms with an R2 of 0.79, NSE of 0.76, MAE of 0.21 (m3/s), MBE of -0.11(m3/s), and RMSE of 0.36 (m3/s). Hybridizing the 2GRU× model with WOA algorithm causes to increase in the value of R2 by 6.8% and reduce in the value of RMSE by 20.4%. Comparatively, standard individual GRU and Bi-GRU models result in an R2 of 0.59 and 0.66, NSE of 0.55 and 0.6, MAE of 0.91 and 0.53 (m3/s), MBE of 0.047 and − 0.06 (m3/s), RMSE of 1.29 and 0.83 (m3/s), respectively.

Multi-model approach for statistical downscaling and bias correction of Earth system model outputs for marine ecosystem impact applications

Abstract Earth system models (ESMs) are the main tool for understanding the impacts of global change and are regularly updated to provide more reliable scenarios of the future. However, their confrontation with observations reveals biases that need to be corrected, especially for impact applications where the absolute scale of the environmental variable is relevant. In addition, marine regional impact studies require fine-scale projections for strategic planning and management actions. Statistical downscaling provides a fast way to produce regional forcings from ESMs and can additionally produce bias-corrected outputs necessary for marine impact applications driven by or fitted to observed data. Statistical downscaling can use different parametric distributions depending on the variables used, and generalised regression can provide a flexible approach for this purpose. We propose a multi-model approach based on non-parametric generalised regression and a set of indicators to select a robust statistical downscaling model that can be used to project future scenarios for marine ecosystems. The empirical cumulative distribution of the variables to be downscaled is modelled, ensuring that not only the mean but also the variance and quantiles (including minima and maxima) are properly represented, improving the prediction of extreme events and taking into account spatial autocorrelation. We incorporated future bias indicators alongside traditional evaluation metrics for model selection to identify and mitigate potential extrapolation errors in future scenario projections, ensuring more robust and plausible downscaled climate outputs. The approach presented here is applied to two contrasted regional case studies, the Bay of Biscay-Celtic Sea ecosystem and the Northern Peru Current ecosystem, using sea surface temperature from the IPSL-CM5A-LR ESM. The results show that a multi-model selection approach is appropriate, as individual model performance is case-specific.

A first-of-its-kind two-body statistical shape model of the arthropathic shoulder: enhancing biomechanics and surgical planning

BackgroundStatistical Shape Models are machine learning tools in computational orthopedics that enable the study of anatomical variability and the creation of synthetic models for pathogenetic analysis and surgical planning. Current models of the glenohumeral joint either describe individual bones or are limited to non-pathologic datasets, failing to capture coupled shape variation in arthropathic anatomy. We aimed to develop a novel combined scapula-proximal-humerus model applicable to clinical populations.MethodsPreoperative computed tomography scans from 45 Reverse Total Shoulder Arthroplasty patients were used to generate three-dimensional models of the scapula and proximal humerus. Correspondence point clouds were combined into a two-body shape model using Principal Component Analysis. Individual scapula-only and proximal-humerus-only shape models were also created for comparison. The models were validated using compactness, specificity, generalization ability, and leave-one-out cross-validation. The modes of variation for each model were also compared. The combined model was described using eigenvector decomposition into single body models. The models were further compared in their ability to predict the shape of one body when given the shape of its counterpart, and the generation of diverse realistic synthetic pairs de novo.ResultsThe scapula and proximal-humerus models performed comparably to previous studies with median average leave-one-out cross-validation errors of 1.08 mm (IQR: 0.359 mm), and 0.521 mm (IQR: 0.111 mm); the combined model was similar with median error of 1.13 mm (IQR: 0.239 mm). The combined model described coupled variations between the shapes equalling 43.2% of their individual variabilities, including the relationship between glenoid and humeral head erosions. The combined model outperformed the individual models generatively with reduced missing shape prediction bias (> 10%) and uniformly diverse shape plausibility (uniformity p-value < .001 vs. .59).ConclusionsThis study developed the first two-body scapulohumeral shape model that captures coupled variations in arthropathic shoulder anatomy and the first proximal-humeral statistical model constructed using a clinical dataset. While single-body models are effective for descriptive tasks, combined models excel in generating joint-level anatomy. This model can be used to augment computational analyses of synthetic populations investigating shoulder biomechanics and surgical planning.

Estimated ultimate recovery prediction of shale gas wells based on stacked integrated learning algorithm

Accurately predicting the estimated ultimate recovery (EUR) of shale gas from a single well is challenging due to geological, engineering, and production factors. Conventional methods often lack sufficient transparency and clarity in the calculation process. As a result, machine learning (ML) algorithms have proven to be an effective alternative. Still, single algorithms are susceptible to outliers or feature selection in the data, leading to unstable predictions. Based on the concept of ensemble learning, this study proposes an intelligent method utilizing automated feature engineering (AutoFE) and stacking ensemble techniques. The method employs Random Forest (RF), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGBoost) as base learners, with Logistic Regression (LR) as the meta-learner. Furthermore, the model is optimized using the Tree-structured Parzen Estimator (TPE) algorithm. The proposed stacking ensemble learning model was validated using a publicly available dataset comprising 506 groups of EUR data of shale gas. The results demonstrate that the proposed stacking ensemble model outperforms individual machine learning models, achieving an R2 of 0.9456, an RMSE of 0.7432, and a MAPE as low as 4.36%. Furthermore, paired t-test results indicate that the use of AutoFE significantly enhances the predictive performance of the model. Furthermore, to enhance the interpretability of the prediction results, the Shapley Additive Explanations (SHAP) technique was employed to conduct an explainable analysis of the machine learning models. This approach revealed the influence trends and magnitudes of reservoir parameters and based learners on the prediction outcomes. The results further indicate that lateral length is the primary factor affecting EUR, followed by proppant loading. This study accurately predicts shale gas EUR and identifies key factors influencing the prediction results, providing valuable insights for predicting shale gas reservoir parameters and optimizing development plans.

Dynamic weighted ensemble model for predictive optimization in green sand casting: Advancing industry 4.0 manufacturing.

Dynamic weighted ensemble model for predictive optimization in green sand casting: Advancing industry 4.0 manufacturing.

Initial seizure episodes risk factors identification during hospitalization of ICU patients: A retrospective analysis of the eICU collaborative research database.

Initial seizure episodes risk factors identification during hospitalization of ICU patients: A retrospective analysis of the eICU collaborative research database.

In vitro evaluation of acetabular host bone coverage and cemented cup stability in total hip arthroplasty.

In vitro evaluation of acetabular host bone coverage and cemented cup stability in total hip arthroplasty.

Detecting Fake News Using Hybrid Machine Learning Models

The increasing diffusion of misinformation in online media has raised alarm as a significant threat to information credibility and societal trust. The ease of disseminating false information across social media platforms, news websites, and digital forums has led to severe consequences, including political manipulation, financial fraud, and public misinformation. This research outlines a robust strategy for detecting misinformation using Natural Language Processing. To take the detection process a step further, the study analyses the implementation of ensemble models. Ensemble learning combines multiple classifiers to improve generalization and robustness, reducing the likelihood of misclassification and helps the model focus on critical words and phrases that contribute to determining the authenticity of news articles. The operational efficacy of the model is measured exploiting standard evaluation assessment metric fidelity, precision, recall, and F1-score to confirm consistent performance. Inclusion of ensemble learning further improves classification accuracy by reducing biases inherent in individual models. Future work in this domain can focus on Live Monitoring for Misinformation, multilingual analysis, and incorporation of context-aware models to further refine detection capabilities. By continuously evolving NLP-based approaches, researchers and technology developers can serve an important function in mitigating the effect of misinformation on social dynamics.

Declining Contribution of Plant Physiological Effects to Global Drought Characteristics With Rising CO2 Using State‐of‐the‐Art Earth System Models

Abstract Vegetation physiology responses to rising atmospheric CO2 can alter the global hydrological cycle, thereby influencing drought occurrence. It has long been controversial and poorly understood how vegetation physiological effects influence meteorological drought characteristics with increasing CO2. To investigate that, we employ multiple CO2 sensitivity experiments of the state‐of‐the‐art Earth System Models (ESMs) in the Coupled Model Intercomparison Project Phase 6 (CMIP6). We quantify drought characteristics in response to rising CO2 using two drought indices: the Standardized Precipitation Index (SPI) and the Standardized Precipitation Evapotranspiration Index (SPEI), with SPEI calculated using both the Penman‐Monteith method (SPEI_PM) and energy‐only method (SPEI_Rn). Our findings reveal that plant physiological effects can robustly induce more intense, frequent, and prolonged droughts under elevated CO2 levels. Spatially, drought intensity as measured by SPI, SPEI_PM, and SPEI_Rn, resulting from CO2 physiological forcing, is projected to increase over 61%, 69%, and 78% of global terrestrial areas, respectively. Notably, we found that the contribution of plant physiological effects () to drought characteristics, including intensity, frequency, and duration, exhibits a significant and spatially extensive declining trend with rising CO2 across most land areas. This declining trend is robustly depicted in both the multi‐model mean and individual models. Vegetation coverage plays an important role in the spatial pattern of . CO2 physiological forcing therefore exerts greater impacts in the tropics, particularly over tropical forests. Our results demonstrate that drought characteristics are expected to become less dependent on plant physiological effects with increasing CO2, a consideration essential for accurate drought projections.

Exploring predictors of clinical response to a transdiagnostic, internet-delivered psychological intervention for people with chronic health conditions.

This study aims to identify participant- and treatment-related characteristics associated with response and deterioration in an internet-delivered intervention for people with chronic health conditions. Understanding these factors is critical for the development and delivery of these treatments. Data were drawn from a randomized controlled trial of a transdiagnostic, internet-delivered intervention for people with chronic health conditions (N = 590). Demographic (e.g., age, gender, employment, education), clinical (e.g., health condition, medication usage, multimorbidity), psychological (e.g., baseline symptom severity), and treatment-related variables were examined. Outcomes included clinically meaningful response (≥ 50% improvement in symptoms) and deterioration (≥ 30% increase in symptoms) in depression, anxiety, and self-reported disability. Multivariable regression models were built to identify significant predictor variables of response and deterioration. Additional indicators of model fit were reported, including the area under the curve and Negelkerke's R². Several predictors of response and deterioration were identified within individual models; however, few predictors were significant across > 1 outcome, while none were consistent across all three outcomes. The final models predicting participant outcomes explained 7%-16% of the variance in the likelihood of response or deterioration. Results suggest that few predictors are meaningfully associated with participants' response or deterioration to treatment. Encouragingly, potentially prognostic variables (e.g., more severe baseline symptoms and multimorbidity) were also nonsignificant. The results support the broad applicability of this treatment approach and for the role of remotely delivered interventions within stepped care frameworks. (PsycInfo Database Record (c) 2025 APA, all rights reserved).

The mechanical response of the condylar process of the mandible under changes of the occlusal plane inclination.

The mechanical response of the condylar process of the mandible under changes of the occlusal plane inclination.

ComBatLS: A Location‐ and Scale‐Preserving Method for Multi‐Site Image Harmonization

ABSTRACTRecent study has leveraged massive datasets and advanced harmonization methods to construct normative models of neuroanatomical features and benchmark individuals' morphology. However, current harmonization tools do not preserve the effects of biological covariates including sex and age on features' variances; this failure may induce error in normative scores, particularly when such factors are distributed unequally across sites. Here, we introduce a new extension of the popular ComBat harmonization method, ComBatLS, that preserves biological variance in features' locations and scales. We use UK Biobank data to show that ComBatLS robustly replicates individuals' normative scores better than other ComBat methods when subjects are assigned to sex‐imbalanced synthetic “sites.” Additionally, we demonstrate that ComBatLS significantly reduces sex biases in normative scores compared to traditional methods. Finally, we show that ComBatLS successfully harmonizes consortium data collected across over 50 studies. R implementation of ComBatLS is available at https://github.com/andy1764/ComBatFamily.

Crop prediction using an enhanced stacked ensemble machine learning model

In India, agriculture is a major sector that fulfils the population's food requirements and significantly contributes to the gross domestic product (GDP). The careful selection of crops is fundamental to maximizing agricultural yield, thereby elevating the economic vitality of the farming community. Precision agriculture (PA) leverages weather and soil data to inform crop selection strategies. Conventional machine learning (ML) models such as decision trees (DT), support vector classifier, K-nearest neighbors (KNN), and extreme gradient boost (XGBoost) have been deployed to predict the best crop. However, these model's efficiency is suboptimal in the current circumstances. The enhanced stacked ensemble ML model is a sophisticated meta-model that addresses these limitations. It harnesses the predictive power of individual ML models, stratified in a layered architecture to improve the prediction accuracy. This advanced model has demonstrated a commendable accuracy rate of 93.1% prediction by taking input of 12 soil parameters such as Nitrogen, Phosphorus, Potassium, and weather parameters such as temperature and rainfall, substantially outperforming the accuracies achieved by the individual contributing models. The efficacy of the proposed meta-model in crop selection based on agronomic parameters signifies a substantial advancement, fortifying the economic resilience of India's agriculture.

Self-adaptive individual tree modeling based on skeleton graph optimization and fractal self-similarity

Self-adaptive individual tree modeling based on skeleton graph optimization and fractal self-similarity

Comparing machine learning methods on Raman spectra from eight different spectrometers.

Comparing machine learning methods on Raman spectra from eight different spectrometers.

Determinants of overall survival in the South African Breast Cancer and HIV Outcomes cohort.

11158 Background: Sub-Saharan Africa (SSA) has very low breast cancer (BC) survival. South Africa, unlike most SSA countries, is an upper middle-income country where patients are less burdened with cancer diagnostic and treatment costs. We aimed to provide five-year overall survival (OS) estimates and determinants among South African women with BC diagnosed and treated within the public health system. Methods: The South African Breast Cancer and HIV Outcomes prospective cohort study enrolled adult women recently diagnosed with invasive BC from four South African academic hospitals. We collected detailed sociodemographic, clinical, treatment, and outcomes data. Women were followed for five years or until December 31, 2023, whichever was earlier, and our primary outcome was five-year OS. Impacts of potential determinants on OS were assessed using individual Cox proportional hazard models focused on nine variable domains ( i.e., treating hospital, social status, BC risk factors, smoking, cardiovascular disease (CVD), HIV status, BC type, BC treatments, and age) and a combined model that also included adjustments for background mortality. Results: Between July 2015 and Feb 2019, we enrolled 2,838 women newly diagnosed with histopathologically confirmed invasive BC, of whom 58% had III or IV disease. At the end of follow-up, 1555 (55%) of the women had died, 1191 (42%) survived for five years, 33 (1%) were censored at study’s end, and 59 (2%) were lost to follow up. The five-year OS was 44.3% (95% CI 42.5-46.2). In the full model, the variables with the largest impact on five-year survival were late-stage at diagnosis (HR 2.4 [95% CI 2.0-2.7] for stage III and HR 5.0 [95% CI 4.1-6.1] for stage IV disease, both compared with stage I/II) and differing degree of treatments received (no treatment received: HR 8.2 [95% CI 5.9-11.3]; chemotherapy and endocrine therapy: HR 1.9 [95% CI 1.4-2.6]; endocrine therapy only: HR 2.2 [95% CI 1.6-3.2]; chemotherapy only: HR 4.8 [95% CI 3.4-6.6]; surgery only: HR 3.0 [95% CI 1.9-4.8]; surgery and chemotherapy: HR 1.8 [95% CI 1.3-2.5], all compared with surgery and endocrine treatments). Other variables significantly associated with survival were treating hospital, relationship status, employment, education, family history of cancer, CVD, and HIV infection. Conclusions: Interventions to improve BC survival in South Africa’s public health system should prioritize earlier diagnosis of cancer and expansion of the infrastructure supporting diagnostic and treatment services. Concurrently, South African BC patients with low socioeconomic status, HIV infection, and/or comorbid CVD are uniquely vulnerable and their barriers to accessing care must be better understood.