Tag
This paper introduces 'alert-list concordance,' a Jaccard-similarity-based metric with patient-level bootstrap confidence intervals, to evaluate how nine imputation strategies affect which ICU patients are flagged for sepsis, finding that pipelines with similar AUROC can produce substantially different alert lists on PhysioNet/CinC 2019 data.
This paper presents a validation protocol for knowledge tracing models, evaluating predictive performance and the stability and faithfulness of explanations using XGBoost and deep learning baselines on ASSISTments datasets.
This study demonstrates that machine learning models using complete blood count (CBC) data outperform C-reactive protein (CRP) alone in differentiating between bacterial and viral infections in pediatric patients.
This paper introduces Solar Intelligence, a hybrid retrieval-augmented framework that unifies solar data analytics, evidence-grounded scientific question answering, and machine learning forecasting to support decision-making in solar energy.
The paper investigates machine learning for train delay prediction in Finland using weather data, demonstrating that domain-informed sensor features improve accuracy over raw observations.
Researchers developed a cost-sensitive XGBoost model to predict ESBL-producing Enterobacteriaceae infections using electronic health record data, aiming to guide empiric antibiotic selection and reduce carbapenem overuse in a 12-hospital study.
This paper proposes a two-stage integrated forecasting model using XGBoost to predict CPU workload in private clouds by first forecasting customer service requests, achieving high accuracy with SMAPE below 7% for most applications.
The paper evaluates how prediction error of nuisance functions relates to causal estimator performance, finding that prediction error is not a consistent measure of causal bias or confidence interval coverage across methods like XGBoost and Double Machine Learning.
XGBoost has been updated to support multi-class and multi-label problems without the need to train separate models for each class or label, introducing the Vector-Leaf Model.
SAGE-XGBoost is a machine learning framework that uses spatially augmented graph embeddings and data augmentation to enhance natural hazard susceptibility mapping under data scarcity, demonstrating superior performance over conventional models.
The paper examines XGBoost's sensitivity to outliers in regression tasks and introduces robust loss functions based on M-, S-, and τ-estimators, proposing MM-XGBoost for a better trade-off between robustness and predictive accuracy.
This preprint compares XGBoost and LSTM for forecasting transmitted heat energy in District Heating Systems, finding that XGBoost consistently outperforms LSTM while offering lower computational cost and environmental impact.
This paper presents a machine learning pipeline for detecting self-introductions in legislative committee testimony, using features like bag-of-words and BERT probabilities. XGBoost achieves the best F1 score of 0.9747, improving further with BERT-augmented features.
Gabriel Chua notes that XGBoost won the KDD 2026 Test of Time Award, recognizing its lasting impact on machine learning.
This paper introduces a leakage-aware evaluation framework for Ethereum actor classification, comparing tree-based models like XGBoost against sequential deep learning models (Transformer, BiLSTM). The authors find that XGBoost outperforms sequence models under leak-reduced conditions while offering lower latency and energy use.
This paper proposes LFS-FRAME, a leakage-free stacked ensemble framework integrating Kolmogorov-Arnold Networks and XGBoost for robust multiclass classification, achieving 89.85% accuracy on major families and 81.74% on sub-families.
This paper presents ClickGuard, a browser extension that detects and spoils clickbait news using a hybrid machine learning model combining transformer embeddings and linguistic features, achieving 91% F1-score on a combined dataset.
This paper develops a multi-dimensional framework to evaluate discrimination, calibration, interpretability, and algorithmic fairness for machine learning-based type 2 diabetes risk prediction models, revealing significant performance degradation under real-world distribution shifts and biases by age and obesity.
This research tests whether Benjamin Graham's classical value investing rules can act as a mathematical 'low-pass filter' to prevent modern machine learning models (XGBoost, AutoGluon) from overfitting to market noise. Using 20 years of S&P 500 data, the authors find that Graham's rules combined with Random Forest achieve high returns with lower risk than complex AI models alone.
This paper presents ClaMPAPP, a hybrid architecture that uses an LLM as an interface to extract features from clinical narratives, which are then passed to an XGBoost classifier for pediatric appendicitis diagnosis, demonstrating improved robustness and safety over end-to-end LLM baselines.