Development Of Machine Learning Models Research Articles

Early prognosis prediction for non-variceal upper gastrointestinal bleeding in the intensive care unit: based on interpretable machine learning

IntroductionThis study aims to construct a mortality prediction model for patients with non-variceal upper gastrointestinal bleeding (NVUGIB) in the intensive care unit (ICU), employing advanced machine learning algorithms. The goal is to identify high-risk populations early, contributing to a deeper understanding of patients with NVUGIB in the ICU.MethodsWe extracted NVUGIB data from the Medical Information Mart for Intensive Care IV (MIMIC-IV, v.2.2) database spanning from 2008 to 2019. Feature selection was conducted through LASSO regression, followed by training models using 11 machine learning methods. The best model was chosen based on the area under the curve (AUC). Subsequently, Shapley additive explanations (SHAP) was employed to elucidate how each factor influenced the model. Finally, a case was randomly selected, and the model was utilized to predict its mortality, demonstrating the practical application of the developed model.ResultsIn total, 2716 patients with NVUGIB were deemed eligible for participation. Following selection, 30 out of a total of 64 clinical parameters collected on day 1 after ICU admission remained associated with prognosis and were utilized for developing machine learning models. Among the 11 constructed models, the Gradient Boosting Decision Tree (GBDT) model demonstrated the best performance, achieving an AUC of 0.853 and an accuracy of 0.839 in the validation cohort. Feature importance analysis highlighted that shock, Glasgow Coma Scale (GCS), renal disease, age, albumin, and alanine aminotransferase (ALP) were the top six features of the GBDT model with the most significant impact. Furthermore, SHAP force analysis illustrated how the constructed model visualized the individualized prediction of death.ConclusionsPatient data from the MIMIC database were leveraged to develop a robust prognostic model for patients with NVUGIB in the ICU. The analysis using SHAP also assisted clinicians in gaining a deeper understanding of the disease.

Developing machine learning models to evaluate the environmental impact of financial policies

The intersection of financial policies and environmental impact is a critical area of research given the urgent need to address climate change and sustainability challenges. This review article explores the current state of machine learning (ML) models used to evaluate the environmental impact of financial policies. We discuss the methodologies, applications, challenges, and future directions of this interdisciplinary field. Emphasis is placed on the integration of economic and environmental data, model interpretability, and the potential of ML to provide actionable insights for policymakers.

Mussel culture monitoring with semi-supervised machine learning on multibeam echosounder data using label spreading

High diversity seabed habitats, such as shellfish aggregations, play a significant role in marine ecosystem sustainability but are susceptible to bottom disturbance induced by anthropogenic activities. Regular monitoring of these habitats with effective mapping methods is therefore essential. Multibeam echosounder (MBES) has been widely used in recent decades for seabed characterization due to its non-destructive manner and extensive spatial coverage compared to traditional methods like bottom sampling. Nevertheless, bottom sampling remains essential to link ground truth with acoustic seabed classification. Using seabed samples and MBES measurements, machine learning techniques are commonly employed to model their relationships and generate classification maps of an extended seabed. However, limited ground truth data, resulting from constraints in regulations, budget, or time, may impede the development of robust machine learning models. To address this challenge, we applied a semi-supervised machine learning method to classify seabed sediments of a blue mussel (Mytilus edulis) cultivation area in the Oosterschelde, the Netherlands. We utilized nine boxcore samples to generate pseudo-labels on MBES data. These pseudo-labels enlarged the training data size, facilitated the training of three comprehensive machine learning algorithms (Gradient Boosting, Random Forest, and Support Vector Machine), and helped to classify the study site into mussel and non-mussel areas. We found the geomorphological and backscatter-related features to be complementary for mussel culture detection. Our classification results were demonstrated effective through expert knowledge of this cultivation area and brought insights for future research on natural mussel habitats.

Evaluating the generalizability of graph neural networks for predicting collision cross section

Ion Mobility coupled with Mass Spectrometry (IM-MS) is a promising analytical technique that enhances molecular characterization by measuring collision cross-section (CCS) values, which are indicative of the molecular size and shape. However, the effective application of CCS values in structural analysis is still constrained by the limited availability of experimental data, necessitating the development of accurate machine learning (ML) models for in silico predictions. In this study, we evaluated state-of-the-art Graph Neural Networks (GNNs), trained to predict CCS values using the largest publicly available dataset to date. Although our results confirm the high accuracy of these models within chemical spaces similar to their training environments, their performance significantly declines when applied to structurally novel regions. This discrepancy raises concerns about the reliability of in silico CCS predictions and underscores the need for releasing further publicly available CCS datasets. To mitigate this, we introduce Mol2CCS which demonstrates how generalization can be partially improved by extending models to account for additional features such as molecular fingerprints, descriptors, and the molecule types. Lastly, we also show how confidence models can support by enhancing the reliability of the CCS estimates.Scientific contributionWe have benchmarked state-of-the-art graph neural networks for predicting collision cross section. Our work highlights the accuracy of these models when trained and predicted in similar chemical spaces, but also how their accuracy drops when evaluated in structurally novel regions. Lastly, we conclude by presenting potential approaches to mitigate this issue.

From Halos to Galaxies. IX. Estimate of Halo Assembly History for SDSS Galaxy Groups

The properties of the galaxies are tightly connected to their host halo mass and halo assembly history. Accurate measurement of the halo assembly history in observation is challenging but crucial to the understanding of galaxy formation and evolution. The stellar-to-halo mass ratio (M */M h) for the centrals has often been used to indicate the halo assembly time t h,50 of the group, where t h,50 is the lookback time at which a halo has assembled half of its present-day virial mass. Using mock data from the semi-analytic models, we find that M */M h shows a significant scatter with t h,50, with a strong systematic difference between the group with a star-forming central (blue group) and passive central (red group). To improve the accuracy, we develop machine learning models to estimate t h,50 for galaxy groups using only observable quantities in the mocks. Since star formation quenching will decouple the co-growth of the dark matter and baryon, we train our models separately for blue and red groups. Our models have successfully recovered t h,50, within an accuracy of ∼1.09 Gyr. With careful calibrations of individual observable quantities in the mocks with Sloan Digital Sky Survey (SDSS) observations, we apply the trained models to the SDSS Yang et al. groups and derive the t h,50 for each group for the first time. The derived SDSS t h,50 distributions are in good agreement with that in the mocks, in particular for blue groups. The derived halo assembly history, together with the halo mass, make an important step forward in studying the halo–galaxy connections in observation.

Analyzing Wav2Vec 1.0 Embeddings for Cross-Database Parkinson's Disease Detection and Speech Features Extraction.

Advancements in deep learning speech representations have facilitated the effective use of extensive unlabeled speech datasets for Parkinson's disease (PD) modeling with minimal annotated data. This study employs the non-fine-tuned wav2vec 1.0 architecture to develop machine learning models for PD speech diagnosis tasks, such as cross-database classification and regression to predict demographic and articulation characteristics. The primary aim is to analyze overlapping components within the embeddings on both classification and regression tasks, investigating whether latent speech representations in PD are shared across models, particularly for related tasks. Firstly, evaluation using three multi-language PD datasets showed that wav2vec accurately detected PD based on speech, outperforming feature extraction using mel-frequency cepstral coefficients in the proposed cross-database classification scenarios. In cross-database scenarios using Italian and English-read texts, wav2vec demonstrated performance comparable to intra-dataset evaluations. We also compared our cross-database findings against those of other related studies. Secondly, wav2vec proved effective in regression, modeling various quantitative speech characteristics related to articulation and aging. Ultimately, subsequent analysis of important features examined the presence of significant overlaps between classification and regression models. The feature importance experiments discovered shared features across trained models, with increased sharing for related tasks, further suggesting that wav2vec contributes to improved generalizability. The study proposes wav2vec embeddings as a next promising step toward a speech-based universal model to assist in the evaluation of PD.

Integrated machine learning models for predictive analysis of thermal and electrical power generation of a photo-thermal system at Catania, Italy

The Photovoltaic-Thermal (PV/T) system is designed for producing both electrical and thermal energy. Its efficiency is improved when temperature-sensitive PV cells are cooled using nanofluid coolants. However, the current models for predicting PV/T system performance with nanofluid cooling are limited. In order to fill this gap, this study aims to develop machine learning models to predict the electrical and thermal efficiencies of a water-based PV/T system. Three types of machine learning algorithms from the supervised learning method were selected in this study because of their ability to handle complex data and provide accurate predictions: the Multi-layer Perceptron (MLP), the Gradient Boosting Regressor (GBR) and the Light Gradient-Boosting Machine (LightGBM). Initially, the models are trained and validated using data consisting of 15,540 samples from a water-based PV/T system. On the basis of the results obtained, the MLP algorithm proved to be the best for the prediction of electrical energy, with an R2 value = 0.9906, against 0.9709 and 0.99 respectively for the other GBR and LGB algorithms, while the LGB algorithm proved effective for the prediction of thermal energy, with an R2 value = 0.983, against 0.983, 0.97 and 0.988 respectively for the GBR and LGB algorithms. Secondly, the best MLP and LGB models previously identified for the water-based PV/T system were adapted to predict, this time, the energy performance of an Ag/water nanofluid-based PV/T system. The results obtained show that these models are highly effective in predicting the electrical and thermal performance of the Ag/water nanofluid PV/T system, even in the absence of real data, making them very useful for real applications.

Estimation of confined compressive strength of LRS‐FRP concrete specimens with computational intelligence

AbstractReinforced concrete structures deteriorate due to changes in temperature, corrosion, and attacks of sulfate and chloride contents. Retrofitting techniques like fiber‐reinforced polymer (FRP) jacketing, known for their strength and corrosion resistance, are increasingly used to strengthen and retrofit deteriorated structural elements. Large rupture strain (LRS)‐FRP composite, composed of polyethylene terephthalate and polyethylene naphthalate, both of which have high tensile strength and high strain at rupture have been used in the studies of many researchers. This research aims to develop a reliable and accurate machine learning (ML) model to estimate the compressive strength of LRS‐FRP confined specimens. A total of 303 LRS‐FRP confined specimens were gathered after a thorough literature review to develop ML models, utilizing the linear regression, support vector regression, regression tree, and artificial neural network (ANN) algorithms. Additionally, 44 analytical models (AMs) were used to compare the performance of the developed ML models. The results revealed that the performance of the developed ANN model was higher among all the ML and AMs. The R‐value and the mean absolute percentage error (MAPE) value of the developed ANN model were 0.9822 and 6.17%, respectively. The sensitivity analysis results show that the height of the specimens had the highest impact followed by the diameter of the specimen, the number of FRP layers and thickness, and then the tensile strength of LRS‐FRP. The ANN‐based mathematical expression is simple and easy to use to predict the compressive strength of the LRS‐FRP strengthened specimens.

Advancing toward precision migraine treatment: Predicting responses to preventive medications with machine learning models based on patient and migraine features.

To develop machine learning models using patient and migraine features that can predict treatment responses to commonly used migraine preventive medications. Currently, there is no accurate way to predict response to migraine preventive medications, and the standard trial-and-error approach is inefficient. In this cohort study, we analyzed data from the Mayo Clinic Headache database prospectively collected from 2001 to December 2023. Adult patients with migraine completed questionnaires during their initial headache consultation to record detailed clinical features and then at each follow-up to track preventive medication changes and monthly headache days. We included patients treated with at least one of the following migraine preventive medications: topiramate, beta-blockers (propranolol, metoprolol, atenolol, nadolol, timolol), tricyclic antidepressants (amitriptyline, nortriptyline), verapamil, gabapentin, onabotulinumtoxinA, and calcitonin gene-related peptide (CGRP) monoclonal antibodies (mAbs) (erenumab, fremanezumab, galcanezumab, eptinezumab). We pre-trained a deep neural network, "TabNet," using 145 variables, then employed TabNet-embedded data to construct prediction models for each medication to predict binary outcomes (responder vs. non-responder). A treatment responder was defined as having at least a 30% reduction in monthly headache days from baseline. All model performances were evaluated, and metrics were reported in the held-out test set (train 85%, test 15%). SHapley Additive exPlanations (SHAP) were conducted to determine variable importance. Our final analysis included 4260 patients. The responder rate for each medication ranged from 28.7% to 34.9%, and the mean time to treatment outcome for each medication ranged from 151.3 to 209.5 days. The CGRP mAb prediction model achieved a high area under the receiver operating characteristics curve (AUC) of 0.825 (95% confidence interval [CI] 0.726, 0.920) and an accuracy of 0.80 (95% CI 0.70, 0.88). The AUCs of prediction models for beta-blockers, tricyclic antidepressants, topiramate, verapamil, gabapentin, and onabotulinumtoxinA were: 0.664 (95% CI 0.579, 0.745), 0.611 (95% CI 0.562, 0.682), 0.605 (95% CI 0.520, 0.688), 0.673 (95% CI 0.569, 0.724), 0.628 (0.533, 0.661), and 0.581 (95% CI 0.550, 0.632), respectively. Baseline monthly headache days, age, body mass index (BMI), duration of migraine attacks, responses to previous medication trials, cranial autonomic symptoms, family history of headache, and migraine attack triggers were among the most important variables across all models. A variable could have different contributions; for example, lower BMI predicts responsiveness to CGRP mAbs and beta-blockers, while higher BMI predicts responsiveness to onabotulinumtoxinA, topiramate, and gabapentin. We developed an accurate prediction model for CGRP mAbs treatment response, leveraging detailed migraine features gathered from a headache questionnaire before starting treatment. Employing the same methods, the model performances for other medications were less impressive, though similar to the machine learning models reported in the literature for other diseases. This may be due to CGRP mAbs being migraine-specific. Incorporating medical comorbidities, genomic, and imaging factors might enhance the model performance. We demonstrated that migraine characteristics are important in predicting treatment responses and identified the most crucial predictors for each of the seven types of preventive medications. Our results suggest that precision migraine treatment is feasible.

Machine learning approaches to detect hepatocyte chromatin alterations from iron oxide nanoparticle exposure

This study focuses on developing machine learning models to detect subtle alterations in hepatocyte chromatin organization due to Iron (II, III) oxide nanoparticle exposure, hypothesizing that exposure will significantly alter chromatin texture. A total of 2000 hepatocyte nuclear regions of interest (ROIs) from mouse liver tissue were analyzed, and for each ROI, 5 different parameters were calculated: Long Run Emphasis, Short Run Emphasis, Run Length Nonuniformity, and 2 wavelet coefficient energies obtained after the discrete wavelet transform. These parameters served as input for supervised machine learning models, specifically random forest and gradient boosting classifiers. The models demonstrated relatively robust performance in distinguishing hepatocyte chromatin structures belonging to the group exposed to IONPs from the controls. The study's findings suggest that iron oxide nanoparticles induce substantial changes in hepatocyte chromatin distribution and underscore the potential of AI techniques in advancing hepatocyte evaluation in physiological and pathological conditions.

Applying Machine Learning for Prescriptive Support: A Use Case with Unfractionated Heparin in Intensive Care Units.

Continuous unfractionated heparin is widely used in intensive care, yet its complex pharmacokinetic properties complicate the determination of appropriate doses. To address this challenge, we developed machine learning models to predict over- and under-dosing, based on anti-Xa results, using a monocentric retrospective dataset. The random forest model achieved a mean AUROC of 0.80 [0.77-0.83], while the XGB model reached a mean AUROC of 0.80 [0.76-0.83]. Feature importance was employed to enhance the interpretability of the model, a critical factor for clinician acceptance. After prospective validation, machine learning models such as those developed in this study could be implemented within a computerized physician order entry (CPOE) as a clinical decision support system (CDSS).

Machine Learning with Clinical and Intraoperative Biosignal Data for Predicting Cardiac Surgery-Associated Acute Kidney Injury.

Early identification of patients at high risk of cardiac surgery-associated acute kidney injury (CSA-AKI) is crucial for its prevention. We aimed to leverage perioperative clinical and intraoperative biosignal data to develop machine learning models that predict CSA-AKI. We introduced a novel approach for extracting relevant features from high-resolution intraoperative biosignals to reflect the patient's baseline status, the extent of unfavorable conditions encountered intraoperatively, and data variability. We developed XGBoost models from 2,003 patients across three consecutive perioperative phases using: 1) only preoperative, 2) pre- and intraoperative, and 3) pre-, intra-, and postoperative variables. The predictive performance progressively improved throughout the three consecutive perioperative phases (e.g., AUROC of 0.767 to 0.797 and 0.840), all surpassing the Thakar Score's performance. According to the SHAP method, intraoperative perfusion pressure was most important in the prediction, highlighting the importance of intraoperative patient management and the use of high-resolution biosignal data in predictive modeling to analyze hemodynamic fluctuations during surgery. Early postoperative biomarkers were also important predictors, highlighting the importance of intensified monitoring early after surgery.

Exploring Prediabetes Pathways Using Explainable AI on Data from Electronic Medical Records.

This study leverages data from a Canadian database of primary care Electronic Medical Records to develop machine learning models predicting type 2 diabetes mellitus (T2D), prediabetes, or normoglycemia. These models are used as a basis for extracting counterfactual explanations and derive personalized changes in biomarkers to prevent T2D onset, particularly in the still reversible prediabetic state. The models achieve satisfactory performance. Furthermore, feature importance analysis underscores the significance of fasting blood sugar and glycated hemoglobin, while counterfactuals explanations emphasize the centrality of keeping body mass index and cholesterol indicators within or close to the clinically desirable ranges. This research highlights the potential of machine learning and counterfactual explanations in guiding preventive interventions that may help slow down the progression from prediabetes to T2D on an individual basis, eventually fostering a recovery from prediabetes to a normoglycemic state.

A Novel Method and Python Library for ECG Signal Quality Assessment.

Electrocardiogram (ECG) is one of the reference cardiovascular diagnostic exams. However, the ECG signal is very prone to being distorted through different sources of artifacts that can later interfere with the diagnostic. For this reason, signal quality assessment (SQA) methods that identify corrupted signals are critical to improve the robustness of automatic ECG diagnostic methods. This work presents a review and open-source implementation of different available indices for SQA as well as introducing an index that considers the ECG as a dynamical system. These indices are then used to develop machine learning models which evaluate the quality of the signals. The proposed index along the designed ML models are shown to improve SQA for ECG signals.

An automatic shoreline extraction method from SAR imagery using DeepLab-v3+ and its versatility

ABSTRACT This study presented an automatic shoreline extraction method from synthetic aperture radar (SAR) imagery using a semantic image segmentation model, DeepLab-v3+. Shorelines extracted from optical satellite images were employed as ground truth. More than 130,000 pairs of labeled images were generated, which substantially enhances the dataset’s objectivity for model training. By using different combination of SAR images from beaches with varying characteristics, several models were constructed. These models were then applied to 15 beaches of Japan to validate their accuracy and versatility. The developed machine learning model demonstrates high accuracy in shoreline extraction, particularly when trained with a diverse set of images from various beaches. This versatility is crucial for the model’s applicability across different coastal environments, underscoring the importance of incorporating a wide range of beach properties into the training dataset.

Exploring Computational Techniques in Preprocessing Neonatal Physiological Signals for Detecting Adverse Outcomes: Scoping Review.

Computational signal preprocessing is a prerequisite for developing data-driven predictive models for clinical decision support. Thus, identifying the best practices that adhere to clinical principles is critical to ensure transparency and reproducibility to drive clinical adoption. It further fosters reproducible, ethical, and reliable conduct of studies. This procedure is also crucial for setting up a software quality management system to ensure regulatory compliance in developing software as a medical device aimed at early preclinical detection of clinical deterioration. This scoping review focuses on the neonatal intensive care unit setting and summarizes the state-of-the-art computational methods used for preprocessing neonatal clinical physiological signals; these signals are used for the development of machine learning models to predict the risk of adverse outcomes. Five databases (PubMed, Web of Science, Scopus, IEEE, and ACM Digital Library) were searched using a combination of keywords and MeSH (Medical Subject Headings) terms. A total of 3585 papers from 2013 to January 2023 were identified based on the defined search terms and inclusion criteria. After removing duplicates, 2994 (83.51%) papers were screened by title and abstract, and 81 (0.03%) were selected for full-text review. Of these, 52 (64%) were eligible for inclusion in the detailed analysis. Of the 52 articles reviewed, 24 (46%) studies focused on diagnostic models, while the remainder (n=28, 54%) focused on prognostic models. The analysis conducted in these studies involved various physiological signals, with electrocardiograms being the most prevalent. Different programming languages were used, with MATLAB and Python being notable. The monitoring and capturing of physiological data used diverse systems, impacting data quality and introducing study heterogeneity. Outcomes of interest included sepsis, apnea, bradycardia, mortality, necrotizing enterocolitis, and hypoxic-ischemic encephalopathy, with some studies analyzing combinations of adverse outcomes. We found a partial or complete lack of transparency in reporting the setting and the methods used for signal preprocessing. This includes reporting methods to handle missing data, segment size for considered analysis, and details regarding the modification of the state-of-the-art methods for physiological signal processing to align with the clinical principles for neonates. Only 7 (13%) of the 52 reviewed studies reported all the recommended preprocessing steps, which could have impacts on the downstream analysis. The review found heterogeneity in the techniques used and inconsistent reporting of parameters and procedures used for preprocessing neonatal physiological signals, which is necessary to confirm adherence to clinical and software quality management system practices, usefulness, and choice of best practices. Enhancing transparency in reporting and standardizing procedures will boost study interpretation and reproducibility and expedite clinical adoption, instilling confidence in the research findings and streamlining the translation of research outcomes into clinical practice, ultimately contributing to the advancement of neonatal care and patient outcomes.

Discovery of an End-to-End Pattern for Contaminant-Oriented Advanced Oxidation Processes Catalyzed by Biochar with Explainable Machine Learning.

The utilization of biochar-catalyzed peroxymonosulfate in advanced oxidation processes (BC-PMS AOPs) is widely acknowledged as an effective and economical method for mitigating emerging contaminants (ECs). Especially, state-of-the-art machine learning (ML) technology has been employed to accurately predict the reaction rate constants of EC degradation in BC-PMS AOPs, primarily focusing on three aspects: performance prediction, operating condition optimization, and mechanism interpretation. However, its real application in specific degradation optimization targeting different ECs is seldom considered, hindering the realization of contaminant-oriented BC-PMS AOPs. Herein, we propose a hierarchical ML pipeline to achieve an end-to-end (E2E) pattern for addressing this issue. First, the overall XGB model, trained with the comprehensive data set, can perform well in predicting the reaction constants of EC degradation in BC-PMS AOPs, additionally providing the basis for further analysis of various ECs. Then, the submodels trained with different EC clusters can offer specific strategies for the selection of the optimum option for BC-PMS AOPs of specific ECs with different HOMO-LUMO gaps, thus forming an E2E operating pattern for BC-PMS AOPs. This study not only increases our understanding of contaminant-oriented optimization of AOPs but also successfully bridges the gap between ML model development and its environmental application.

Classification of Latilactobacillus sakei subspecies based on MALDI-TOF MS protein profiles using machine learning models.

Latilactobacillus sakei plays a significant role in the realm of food bacteria. One particular subspecies of L. sakei is employed as a protective agent during food fermentation, whereas another strain is responsible for food spoilage. Hence, it is crucial to precisely differentiate between the two subspecies of L. sakei. In this study, machine learning models based on protein mass peaks were developed for the first time to distinguish L. sakei subspecies. Furthermore, the efficacy of three commonly used machine learning algorithms for microbial classification was evaluated. Our results provide the foundation for future research on developing machine learning models for the classification of microbial species or subspecies. In addition, the developed model can be used in the food industry to monitor L. sakei subspecies in fermented foods in a time- and cost-effective method for food quality and safety.

A novel framework for crash frequency prediction: Geographic support vector regression based on agent-based activity models in Greater Melbourne

The field of spatial analysis in traffic crash studies can often enhance predictive performance by addressing the inherent spatial dependence and heterogeneity in crash data. This research introduces the Geographical Support Vector Regression (GSVR) framework, which incorporates generated distance matrices, to assess spatial variations and evaluate the influence of a wide range of factors, including traffic, infrastructure, socio-demographic, travel demand, and land use, on the incidence of total and fatal-or-serious injury (FSI) crashes across Greater Melbourne’s zones. Utilizing data from the Melbourne Activity-Based Model (MABM), the study examines 50 indicators related to peak hour traffic and various commuting modes, offering a detailed analysis of the multifaceted factors affecting road safety. The study shows that active transportation modes such as walking and cycling emerge as significant indicators, reflecting a disparity in safety that heightens the vulnerability of these road users. In contrast, car commuting, while a consistent factor in crash risks, has a comparatively lower impact, pointing to an inherent imbalance in the road environment. This could be interpreted as an unequal distribution of risk and safety measures among different types of road users, where the infrastructure and policies may not adequately address the needs and vulnerabilities of pedestrians and cyclists compared to those of car drivers. Public transportation generally offers safer travel, yet associated risks near train stations and tram stops in city center areas cannot be overlooked. Tram stops profoundly affect total crashes in these areas, while intersection counts more significantly impact FSI crashes in the broader metropolitan area. The study also uncovers the contrasting roles of land use mix in influencing FSI versus total crashes. The proposed framework presents an approach for dynamically extracting distance matrices of varying sizes tailored to the specific dataset, providing a fresh method to incorporate spatial impacts into the development of machine learning models. Additionally, the framework extends a feature selection technique to enhance machine learning models that typically lack comprehensive feature selection capabilities.

Prediction of 28-Day All-Cause Mortality in Heart Failure Patients with Clostridioides difficile Infection Using Machine Learning Models: Evidence from the MIMIC-IV Database.

Heart failure (HF) may induce bowel hypoperfusion, leading to hypoxia of the villa of the bowel wall and the occurrence of Clostridioides difficile infection (CDI). However, the risk factors for the development of CDI in HF patients have yet to be fully illustrated, especially because of a lack of evidence from real-world data. Clinical data and survival situations of HF patients with CDI admitted to ICU were extracted from the Medical Information Mart for Intensive Care (MIMIC)-IV database. For developing a model that can predict 28-day all-cause mortality in HF patients with CDI, the Recursive Feature Elimination with Cross-Validation (RFE-CV) method was used for feature selection. And nine machine learning (ML) algorithms, including logistic regression (LR), decision tree, Bayesian, adaptive boosting, random forest (RF), gradient boosting decision tree, XGBoost, light gradient boosting machine, and categorical boosting, were applied for model construction. After training and hyperparameter optimization of the models through grid search 5-fold cross-validation, the performance of models was evaluated by the area under curve (AUC), accuracy, sensitivity, specificity, precision, negative predictive value, and F1 score. Furthermore, the SHapley Additive exPlanations (SHAP) method was used to interpret the optimal model. A total of 526 HF patients with CDI were included in the study, of whom 99 cases (18.8%) experienced death within 28 days. Eighteen of the 57 variables were selected for the model construction algorithm for model construction. Among the ML models considered, the RF model emerged as the optimal model achieving the accuracy, F1-score, and AUC values of 0.821, 0.596, and 0.864, respectively. The net benefit of the model surpassed other models at 16%-22% threshold probabilities based on decision curve analysis. According to the importance of features in the RF model, red blood cell distribution width, blood urea nitrogen, Simplified Acute Physiology Score II, Sequential Organ Failure Assessment, and white blood cell count were highlighted as the five most influential variables. We developed ML models to predict 28-day all-cause mortality in HF patients associated with CDI in the ICU, which are more effective than the conventional LR model. The RF model has the best performance among all the ML models employed. It may be useful to help clinicians identify high-risk HF patients with CDI.