Gradient Boosting Research Articles

Characterizing Chromophoric Dissolved Organic Matter Spatio-Temporal Variability in North Andean Patagonian Lakes Using Remote Sensing Information and Environmental Analysis

Chromophoric dissolved organic matter (CDOM) is crucial in aquatic ecosystems, influencing light penetration and biogeochemical processes. This study investigates the CDOM variability in seven oligotrophic lakes of North Andean Patagonia using Landsat 8 imagery. An empirical band ratio model was calibrated and validated for the estimation of CDOM concentrations in surface lake water as the absorption coefficient at 440 nm (acdom440, m−1). Of the five atmospheric corrections evaluated, the QUAC (Quick Atmospheric Correction) method demonstrated the highest accuracy for the remote estimation of CDOM. The application of separate models for deep and shallow lakes yielded superior results compared to a combined model, with R2 values of 0.76 and 0.82 and mean absolute percentage errors (MAPEs) of 14% and 22% for deep and shallow lakes, respectively. The spatio-temporal variability of CDOM was characterized over a five-year period using satellite-derived acdom440 values. CDOM concentrations varied widely, with very low values in deep lakes and moderate values in shallow lakes. Additionally, significant seasonal fluctuations were evident. Lower CDOM concentrations were observed during the summer to early autumn period, while higher concentrations were observed in the winter to spring period. A gradient boosting regression tree analysis revealed that inter-lake differences were primarily influenced by the lake perimeter to lake area ratio, mean lake depth, and watershed area to lake volume ratio. However, seasonal CDOM variation was largely influenced by Lake Nahuel Huapi water storage (a proxy for water level variability at a regional scale), followed by precipitation, air temperature, and wind. This research presents a robust method for estimating low to moderate CDOM concentrations, improving environmental monitoring of North Andean Patagonian Lake ecosystems. The results deepen the understanding of CDOM dynamics in low-impact lakes and its main environmental drivers, enhance the ability to estimate lacustrine carbon stocks on a regional scale, and help to predict the effects of climate change on this important variable.

Predicting Outcomes after Hospitalisation for COPD Exacerbation Using Machine Learning

BackgroundEarly readmission and death are critical adverse outcomes following hospitalisation due to exacerbation of chronic obstructive pulmonary disease (ECOPD). This study aimed to develop and validate machine learning models to enhance the prediction of these outcomes after ECOPD hospitalisation.MethodsUtilising a nationwide database, data from the index ECOPD hospitalisation and the preceding year were collected. Prediction models for 30-day readmission and death were developed using logistic lasso regression, random forest, extreme gradient boosting (XGBoost), and neural network, with the LACE index serving as a reference. Model performance was assessed with receiver operating characteristic (ROC) curves and calibration plots from the validation dataset. Key predictors were identified using SHapley Additive exPlanations.ResultsThe study included 101 011 hospitalisations in the development dataset and 17 565 in the validation dataset. The rates of 30-day readmission and death were 29.1% and 4.3%, respectively. AmoXGBoost outperformed other models, achieving an area under the ROC curve of 0.721 (95% CI, 0.713–0.729) for readmission and 0.809 (95% CI, 0.794–0.824) for death, both exceeding the corresponding values for the LACE index (0.651 and 0.641). All machine learning models demonstrated good calibration. The number of hospitalisations in the previous year and the lowest haemoglobin level during the index hospitalisation were the top predictors of readmission and death, respectively.ConclusionsApplying machine learning techniques to large-scale data effectively improves the prediction of early readmission and death following ECOPD hospitalisation. Identifying critical prognostic factors could enhance targeted post-discharge care for this high-risk patient group.

Predicting the solubility of CO2 and N2 in ionic liquids based on COSMO-RS and machine learning

As ionic liquids (ILs) continue to be prepared, there is a growing need to develop theoretical methods for predicting the properties of ILs, such as gas solubility. In this work, different strategies were employed to obtain the solubility of CO2 and N2, where a conductor-like screening model for real solvents (COSMO-RS) was used as the basis. First, experimental data on the solubility of CO2 and N2 in ILs were collected. Then, the solubility of CO2 and N2 in ILs was predicted using COSMO-RS based on the structures of cations, anions, and gases. To further improve the performance of COSMO-RS, two options were used, i.e., the polynomial expression to correct the COSMO-RS results and the combination of COSMO-RS and machine learning algorithms (eXtreme Gradient Boosting, XGBoost) to develop a hybrid model. The results show that the COSMO-RS with correction can significantly improve the prediction of CO2 solubility, and the corresponding average absolute relative deviation (AARD) is decreased from 43.4% to 11.9%. In contrast, such an option cannot improve that of the N2 dataset. Instead, the results obtained from coupling machine learning algorithms with the COSMO-RS model agree well with the experimental results, with an AARD of 0.94% for the solubility of CO2 and an average absolute deviation (AAD) of 0.15% for the solubility of N2.

Association between nutritional status indices and non-alcoholic fatty liver disease in older adults: insights from the National Health and Nutrition Examination Survey 2017-2018.

While previous studies have identified a relationship between dietary intake and the risk of non-alcoholic fatty liver disease (NAFLD), the influence of overall nutritional status on NAFLD development has not been thoroughly investigated. This study sought to explore the association between different nutritional status indicators and NAFLD among the older adults. Nutritional status was evaluated using controlling nutritional status (CONUT), prognostic nutritional index (PNI) and nutritional risk index (GNRI), while NAFLD was identified based on a controlled attenuation parameter ≥ 285 dB/m, measured using transient elastography. The analysis included multivariate regression, receiver operating characteristic analysis, eXtreme Gradient Boosting and subgroup analysis to investigate the relationships between nutritional status indices and NAFLD. The study enrolled 1409 participants for the main analysis, with an NAFLD prevalence of 44·7 %. After accounting for potential confounders, a positive association between PNI and NAFLD was observed. Participants in the third and fourth quartiles of PNI showed increased odds of NAFLD compared with the lowest quartile (Q3: OR = 1·45, 95 % CI (1·03, 2·05); Q4: OR = 2·27, 95 % CI (1·59, 3·24)). Similarly, higher GNRI quartiles were significantly associated with greater odds of NAFLD (Q4 v. Q1: aOR = 1·84; 95 % CI (1·28, 2·65)). Conversely, higher CONUT values were linked to a reduced prevalence of NAFLD (OR = 0·65, 95 % CI (0·48, 0·87)). This study highlights that suboptimal nutritional status, indicating overnutrition as evaluated by PNI, GNRI and CONUT, is positively linked with the risk of NAFLD in individuals aged 50 years and above.

Machine Learning based Pointing Corrections for Radio/Sub-millimeter Telescopes

Radio, sub-millimeter and millimeter ground-based telescopes are powerful instruments for studying the gas and dust-rich regions of the Universe that are invisible at optical wavelengths, but the pointing accuracy is crucial for obtaining high-quality data. Pointing errors are small deviations of the telescope’s orientation from its desired direction. The telescopes use linear regression pointing models to correct for these errors, taking into account various factors such as weather conditions, telescope mechanical structure, and the target’s position in the sky. However, residual pointing errors can still occur due to factors that are hard to model accurately, such as thermal and gravitational deformation and environmental conditions like humidity and wind. Here we present a proof-of-concept for reducing pointing error for the Atacama Pathfinder EXperiment (APEX) telescope in the high-altitude Atacama Desert in Chile based on machine learning. Using historic pointing data from 2022, we trained eXtreme Gradient Boosting (XGBoost) models that reduced the root-mean-square errors (RMSE) for azimuth and elevation (horizontal and vertical angle) pointing corrections by 4.3% and 9.5%, respectively, on hold-out test data. Our results will inform operations of current and future facilities such as the next-generation Atacama Large Aperture Submillimeter Telescope (AtLAST)

Development of a Practical Prediction Model for Adverse Neonatal Outcomes at the Start of the Second Stage of Labor.

To develop a prediction model for adverse neonatal outcomes using electronic fetal monitoring (EFM) interpretation data and other relevant clinical information known at the start of the second stage of labor. This was a retrospective cohort study of individuals who labored and delivered at two academic medical centers between July 2016 and June 2020. Individuals were included if they had a singleton gestation at term (more than 37 weeks of gestation), a vertex-presenting, nonanomalous fetus, and planned vaginal delivery and reached the start of the second stage of labor. The primary outcome was a composite of severe adverse neonatal outcomes. We developed and compared three modeling approaches to predict the primary outcome using factors related to EFM data (as interpreted and entered in structured data fields in the electronic health record by the bedside nurse), maternal comorbidities, and labor characteristics: traditional logistic regression, LASSO (least absolute shrinkage and selection operator), and extreme gradient boosting. Model discrimination and calibration were compared. Predicted probabilities were stratified into risk groups to facilitate clinical interpretation, and positive predictive values for adverse neonatal outcomes were calculated for each. A total of 22,454 patients were included: 14,820 in the training set and 7,634 in the test set. The composite adverse neonatal outcome occurred in 3.2% of deliveries. Of the three modeling methods compared, the logistic regression model had the highest discrimination (0.690, 95% CI, 0.656-0.724) and was well calibrated. When stratified into risk groups (no increased risk, higher risk, and highest risk), the rates of the composite adverse neonatal outcome were 2.6% (95% CI, 2.3-3.1%), 6.7% (95% CI, 4.6-9.6%), and 10.3% (95% CI, 7.6-13.8%), respectively. Factors with the strongest associations with the composite adverse neonatal outcome included the presence of meconium (adjusted odds ratio [aOR] 2.10, 95% CI, 1.68-2.62), fetal tachycardia within the 2 hours preceding the start of the second stage (aOR 1.94, 95% CI, 1.03-3.65), and number of prior deliveries (aOR 0.77, 95% CI, 0.60-0.99).

Fatigue Life Prediction of 2024-T3 Al Alloy by Integrating Particle Swarm Optimization—Extreme Gradient Boosting and Physical Model

The multi-parameter characteristics of the physical model pose a challenge to the fatigue life prediction of 2024-T3 aluminum (Al) alloy. In response to this issue, a parameter-solving method that integrates particle swarm optimization (PSO) with extreme gradient boosting (XGBoost) is proposed in this study. The fatigue performance and failure mechanism of the 2024-T3 Al alloy are analyzed. Furthermore, the fatigue life prediction physical model of the 2024-T3 Al alloy is established by using the energy method of fracture mechanics. The physical model incorporates critical physical parameters. Meanwhile, the PSO algorithm optimizes the hyperparameters of the XGBoost model based on fatigue data of the 2024-T3 Al alloy. Eventually, the optimized XGBoost model is used to solve the parameters of the physical model. Furthermore, the analytical equation of the fatigue life prediction model is obtained. This paper provides a new method for solving the parameters of the fatigue life prediction model, which reduces the error and cost of obtaining the model parameters in the experiment and shortens the time required.

Effective Machine Learning Techniques for Dealing with Poor Credit Data

Credit risk is a crucial component of daily financial services operations; it measures the likelihood that a borrower will default on a loan, incurring an economic loss. By analysing historical data for assessment of the creditworthiness of a borrower, lenders can reduce credit risk. Data are vital at the core of the credit decision-making processes. Decision-making depends heavily on accurate, complete data, and failure to harness high-quality data would impact credit lenders when assessing the loan applicants’ risk profiles. In this paper, an empirical comparison of the robustness of seven machine learning algorithms to credit risk, namely support vector machines (SVMs), naïve base, decision trees (DT), random forest (RF), gradient boosting (GB), K-nearest neighbour (K-NN), and logistic regression (LR), is carried out using the Lending Club credit data from Kaggle. This task uses seven performance measures, including the F1 Score (recall, accuracy, and precision), ROC-AUC, and HL and MCC metrics. Then, the harnessing of generative adversarial networks (GANs) simulation to enhance the robustness of the single machine learning classifiers for predicting credit risk is proposed. The results show that when GANs imputation is incorporated, the decision tree is the best-performing classifier with an accuracy rate of 93.01%, followed by random forest (92.92%), gradient boosting (92.33%), support vector machine (90.83%), logistic regression (90.76%), and naïve Bayes (89.29%), respectively. The classifier is the worst-performing method with a k-NN (88.68%) accuracy rate. Subsequently, when GANs are optimised, the accuracy rate of the naïve Bayes classifier improves significantly to (90%) accuracy rate. Additionally, the average error rate for these classifiers is over 9%, which implies that the estimates are not far from the actual values. In summary, most individual classifiers are more robust to missing data when GANs are used as an imputation technique. The differences in performance of all seven machine learning algorithms are significant at the 95% level.

A SYSTEMATIC LITERATURE REVIEW OF PREDICTIVE MODELS AND ANALYTICS IN AI-DRIVEN CREDIT SCORING

This systematic review examines the transformative role of AI-driven models in credit scoring, highlighting their advances over traditional statistical methods in terms of predictive accuracy, adaptability, and inclusivity. By synthesizing findings from 70 studies, this review demonstrates that machine learning techniques, particularly ensemble models such as random forests and gradient boosting, effectively capture complex, non-linear relationships in borrower data, providing more accurate risk assessments across diverse demographics. Deep learning models, especially convolutional and recurrent neural networks, extend credit scoring capabilities to unstructured and alternative data sources, supporting financial inclusion by enabling assessments of individuals without traditional credit histories. Hybrid models that integrate logistic regression with neural networks offer an optimal balance between interpretability and predictive power, addressing regulatory demands for transparency while maintaining robust accuracy. Ensemble techniques like stacking and blending enhance model adaptability, allowing credit scoring systems to integrate multiple perspectives and improve prediction accuracy in varied borrower contexts. Despite these advancements, challenges remain in the form of ethical concerns and the need for model interpretability, particularly with complex deep-learning architectures. The review underscores the importance of developing fairness-aware and explainable AI frameworks to ensure that as AI-driven credit scoring evolves, it remains both transparent and equitable. These insights suggest that with careful attention to ethics and transparency, AI has the potential to create a more inclusive and resilient credit scoring landscape, accommodating the needs of an increasingly diverse global population.

Machine Learning Models Leveraging Smartphone-Based Patient Mobility Data Can Accurately Predict Functional Outcomes After Spine Surgery

Objective: The development of adjacent segment disease or the progression of spondylosis following the surgical treatment of spinal stenosis and spondylolisthesis is well documented and can lead to subsequent functional decline after a successful index surgery. The early detection of negative inflection points during patients’ functional recovery can improve timely intervention. In this study, we developed machine learning (ML) models to predict the occurrence of post-operative decline in patient mobility. Methods: Patients receiving spine surgery for degenerative spinal stenosis or spondylolisthesis were retroactively consented and enrolled. Activity data (steps-per-day) previously recorded across a 4-year peri-operative were collected alongside relevant clinical and demographic variables. Logistic regression (LR), random forest (RF), and extreme gradient boosting (XGBoost) ML models were constructed and trained on 80% of the dataset and validated using the remaining 20%. The study’s primary endpoint was the models’ ability to predict post-operative decline in patient mobility. Results: A total of 75 patients were included. Following training, RF and XGBoost models achieved accuracy values of 86.7% (sensitivity 80%, specificity 90%) and 80% (sensitivity 60%, specificity 90%), respectively, in predicting post-operative functional decline. The LR model was the least effective with an accuracy of 73.3% (sensitivity 50%, specificity 88.8%). Receiver operating characteristic curves showed an area under the curve of 0.80 for RF, 0.70 for XGBoost, and 0.69 for LR. Conclusions: ML models trained on activity data collected from smartphones successfully forecast functional decline in post-operative activity following spine surgery. These results lay the groundwork for the future integration of ML into the surgeon’s toolbox for prognostication and surgical planning.

Optimized bp Neural Network Based on Improved Dung Beetle Optimization Algorithm to Predict High-Performance Concrete Compressive Strength

In modern architecture, the structural safety of buildings largely depends on the compressive strength of high-performance concrete (HPC), which is determined by the complex nonlinear relationships between its components. In order to more accurately forecast HPC’s compressive strength, this paper proposes a prediction model based on an improved dung beetle optimization algorithm (OTDBO)-optimized backpropagation neural network (BPNN). Extreme Gradient Boosting (XGBoost) is employed to determine the inputs for the BPNN, enhancing the computational efficiency under high-dimensional data feature conditions. To address the issues of local optima entrapment and slow convergence in the dung beetle optimization algorithm (DBO), four improvements were made to enhance its performance. In the initial population generation stage, the optimal Latin hypercube method was used to increase the population diversity. In the rolling stage, the osprey optimization algorithm’s global exploration strategy was introduced to improve the global search capability. The variable spiral search strategy was employed in the reproduction stage, and an adaptive t-distribution perturbation strategy was combined in the foraging stage to enhance the algorithm’s adaptability and search efficiency. The improved dung beetle optimization algorithm (OTDBO) outperformed other algorithms in performance tests on the CEC2017 benchmark functions. In terms of predicting the compressive strength of HPC, the XG-OTDBO-BP model developed in this study outperformed models optimized by other algorithms in terms of fitting outcomes and prediction accuracy. These findings support the XG-OTDBO-BP model’s superiority in the compressive strength of HPC prediction.

Data-Driven Optimised XGBoost for Predicting the Performance of Axial Load Bearing Capacity of Fully Cementitious Grouted Rock Bolting Systems

This article investigates the application of eXtreme gradient boosting (XGBoost) and hybrid metaheuristics optimisation techniques to predict the axial load bearing capacity of fully grouted rock bolting systems. For this purpose, a comprehensive dataset of 72 pull-out tests was built, considering various influential parameters such as three water-to-grout (W/G) ratios, five ranges of curing time (CT), three different grout admixtures with two different fly ash (FA) contents, and two different diameter confinements (DCs). Additionally, to find out the effect of the mechanical behaviour of grouts on the performance of fully grouted rock bolting systems, seventy-two uniaxial compression strength (UCS) samples were cast and tested simultaneously with pull-out samples. The UCS samples were prepared with the same details as the pull-out samples to avoid any inconsistency. The results highlight that peak load values generally increase with longer curing times, lower W/G, and higher UCS and DC values. The main novelty of this paper lies in its data-driven approach, using various XGBoost models. This method offers a time-, cost-, and labour-efficient alternative to traditional experimental methods for predicting rock bolt performance. For this purpose, after building the dataset and dividing it randomly into two training and testing datasets, five different XGBoost models were developed: a standalone XGBoost model and four hybrid models incorporating Harris hawk optimisation (HHO), the jellyfish search optimiser (JSO), the dragonfly algorithm (DA), and the firefly algorithm (FA). These models were subsequently evaluated for their ability to predict peak load values. The results demonstrate that all models effectively predicted peak load values, but the XGBoost-JSO hybrid model demonstrated superior performance, achieving the highest R-squared coefficients of 0.987 and 0.988 for the training and testing datasets, respectively. Sensitivity analysis revealed that UCS values were the most influential parameter, while FA content had the least impact on the maximum peak load values of fully cementitious grouted rock bolts.

Policy-driven innovations in fraud prevention: Developing an ARFLGB-XGBoost early warning model to mitigate online romance scams in telecommunication networks

To address the escalating online romance scams within telecom fraud, we developed an Adaptive Random Forest Light Gradient Boosting (ARFLGB)-XGBoost early warning system. Our method involves compiling detailed Online Romance Scams (ORS) incident data into a 24-variable dataset, categorized to analyze feature importance with Random Forest and LightGBM models. An innovative adaptive algorithm, the Adaptive Random Forest Light Gradient Boosting, optimizes these features for integration with XGBoost, enhancing early Online romance scams threat detection. Our model showed significant performance improvements over traditional models, with accuracy gains of 3.9%, a 12.5% increase in precision, recall improvement by 5%, an F1 score increase by 5.6%, and a 5.2% increase in Area Under the Curve (AUC). This research highlights the essential role of advanced fraud detection in preserving communication network integrity, contributing to a stable economy and public safety, with implications for policymakers and industry in advancing secure communication infrastructure.

Determination of creep function coefficients of viscoelastic pipes using a transient-guided machine learning model

ABSTRACT This study introduces a novel framework for the estimation of creep function coefficients in viscoelastic pipelines, employing a combination of machine learning (ML) and transient hydraulics. Specifically, the efficacy of eXtreme Gradient Boosting (XGBoost) as an advanced regression method is evaluated within this framework. A transient simulation model, utilizing the method of characteristics, is formulated to create the reservoir-pipe-valve (RPV) scenarios under diverse boundary conditions. After conducting model calibration, the model is used to generate datasets with the transient hydraulic responses at the measurement points. The fast Fourier transform (FFT) is then applied to transform the generated samples into the frequency domain. Feature selection is accomplished through principal component analysis (PCA) to identify optimal input variables for XGBoost. In estimating the creep function, six coefficients are employed, with the feature selection analysis indicating that each coefficient is associated with a specific signal. Importantly, it is demonstrated that attempting to estimate all six coefficients using a single set of signals is unfeasible. The results affirm the accuracy of the proposed ML-based framework in determining creep function coefficients.

Multisource Data-Driven Carbon Price Composite Forecasting Model: Based on Feature Selection and Multiscale Prediction Strategy

Accurate prediction of carbon trading prices is crucial for guiding investors to make informed decisions and assisting governments in formulating scientific carbon trading policies. This paper introduces a multisource data-driven carbon price composite forecasting model, aimed at enhancing prediction accuracy through advanced data processing and analysis methods. The model initially employs a secondary decomposition strategy, including Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and Variational Mode Decomposition (VMD) methods, to decompose the original data series into three sub-sequences of different frequencies. Subsequently, it utilizes the Partial Autocorrelation Function (PACF) and Random Forest algorithm for feature selection to determine the input variables for different frequency sequences and conducts in-depth analysis and selection of influencing factors, including unstructured data. Furthermore, the model employs a multiscale forecasting strategy, combining Particle Swarm Optimization (PSO) enhanced Bidirectional Long Short-Term Memory (BiLSTM) and Extreme Gradient Boosting (XGBoost) models, along with the Autoregressive Integrated Moving Average (ARIMA) method, to predict based on the characteristics of different frequency components. Finally, the forecasts are integrated using PSO-BiLSTM to form a comprehensive forecast. Notably, given the high correlation between the trend series and influencing factor variables, the model jointly predicts them. A case study based on the Guangdong carbon market in China demonstrates that the proposed composite forecasting model outperforms other benchmark models, with Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE) values of 0.4009, 0.2699, and 0.5183%, respectively. This forecasting model provides an effective tool for predicting and analyzing carbon price fluctuations, offering new insights for precise carbon market price predictions.

Deep Learning for Opportunistic Rain Estimation via Satellite Microwave Links

Accurate precipitation measurement is critical for managing flood and drought risks. Traditional meteorological tools, such as rain gauges and remote sensors, have limitations in resolution, coverage, and cost-effectiveness. Recently, the opportunistic use of microwave communication signals has been explored to improve precipitation estimation. While there is growing interest in using satellite-to-earth microwave link (SMLs) for machine learning-based precipitation estimation, direct rainfall estimation from raw signal-to-noise ratio (SNR) data via deep learning remains underexplored. This study investigates a range of machine learning (ML) approaches, including deep learning (DL) models and traditional methods like gradient boosting machine (GBM), for estimating rainfall rates from SNR data collected by interactive satellite receivers. We develop real-time models for rainfall detection and estimation using downlink SNR signals from satellites to user terminals. By leveraging a year-long dataset from multiple locations—including SNR measurements paired with disdrometer and rain-gauge data—we explore and evaluate various ML models. Our final models include ensemble approaches for both rainfall detection and cumulative rainfall estimation. The proposed models provide a reliable solution for estimating precipitation using Earth–satellite microwave links, potentially improving precipitation monitoring. Compared to the state-of-the-art power-law-based models applied to similar datasets reported in the literature, our ML models achieve a 46% reduction in the RMSE

Identifying the spatial pattern and driving factors of nitrate in groundwater using a novel framework of interpretable stacking ensemble learning

Groundwater nitrate contamination poses a potential threat to human health and environmental safety globally. This study proposes an interpretable stacking ensemble learning (SEL) framework for enhancing and interpreting groundwater nitrate spatial predictions by integrating the two-level heterogeneous SEL model and SHapley Additive exPlanations (SHAP). In the SEL model, five commonly used machine learning models were utilized as base models (gradient boosting decision tree, extreme gradient boosting, random forest, extremely randomized trees, and k-nearest neighbor), whose outputs were taken as input data for the meta-model. When applied to the agricultural intensive area, the Eden Valley in the UK, the SEL model outperformed the individual models in predictive performance and generalization ability. It reveals a mean groundwater nitrate level of 2.22 mg/L-N, with 2.46% of sandstone aquifers exceeding the drinking standard of 11.3 mg/L-N. Alarmingly, 8.74% of areas with high groundwater nitrate remain outside the designated nitrate vulnerable zones. Moreover, SHAP identified that transmissivity, baseflow index, hydraulic conductivity, the percentage of arable land, and the C:N ratio in the soil were the top five key driving factors of groundwater nitrate. With nitrate threatening groundwater globally, this study presents a high-accuracy, interpretable, and flexible modeling framework that enhances our understanding of the mechanisms behind groundwater nitrate contamination. It implies that the interpretable SEL framework has great promise for providing valuable evidence for environmental management, water resource protection, and sustainable development, particularly in the data-scarce area.

A Predictive Model for Perinatal Brain Injury Using Machine Learning Based on Early Birth Data

Background/Objective: It is difficult to predict perinatal brain injury, and performing brain magnetic resonance imaging (MRI) based on suspected injury remains a clinical challenge. Therefore, we aimed to develop a reliable method for predicting perinatal brain injury using a machine learning model with early birth data. Methods: Neonates admitted to our institution from January 2017 to June 2024 with a gestational age of ≥36 weeks, a birth weight of ≥1800 g, admission within 6 h of birth, and who underwent brain MRI to confirm perinatal brain injury were included. Various machine learning models, including gradient boosting, were trained using early birth data to predict perinatal brain injury. Synthetic minority over-sampling and adaptive synthetic sampling (ADASYN) were applied to address class imbalance. Model performance was evaluated using accuracy, F1 score, and ROC curves. Feature importance scores and Shapley additive explanations (SHAP) values were also calculated. Results: Among 179 neonates, 39 had perinatal brain injury. There were significant differences between the injury and non-injury groups in mode of delivery, Apgar scores, capillary pH, lactate dehydrogenase (LDH) levels, and whether therapeutic hypothermia was performed. The gradient boosting model with the ADASYN method achieved the best performance. In terms of feature importance scores, the 1 min Apgar score was the most influential predictor. Additionally, SHAP analysis showed that LDH levels had the highest SHAP values. Conclusion: the gradient boosting model with ADASYN oversampling effectively predicts perinatal brain injury, potentially improving early detection for predicting long-term outcomes, reducing unnecessary MRI scans, and lowering healthcare costs.

Experimental Study and Machine Learning-Based Prediction of the Abrasion Resistance of Manufactured Sand Concrete

To systematically analyze the impact of manufactured sand on the abrasion resistance of concrete, this paper investigates the correlation between sand type, sand ratio, stone powder content, compressive strength, and the abrasion resistance of manufactured sand concrete. Grey correlation analysis was conducted to assess the impact priority of each factor affecting the abrasion resistance, and prediction models for the abrasion resistance were developed using XGBoost, random forest, AdaBoost, and gradient boosting. The results indicate that compared to river sand concrete, C30 and C40 concrete prepared with limestone and diabase manufactured sand has 20% higher abrasion resistance due to the presence of stone powder and higher roughness and solidity. Within the range of 0.40 to 0.44, a lower sand ratio leads to higher abrasion resistance. For concrete prepared with manufactured sand containing 5% to 11% stone powder, the best abrasion resistance can be attained at a stone powder content of 9%, and microscopic analysis suggests the highest concrete density at this level. According to grey system theory, the influence of each affecting factor on the abrasion resistance follows the order: sand ratio > crushing value > roughness > compressive strength > stone powder content > 0.6. Compared to gradient boosting, random forest, and AdaBoost models, the XGBoost model adopted in this study showed relatively higher R2 and lower RMSE in both the training and testing sets, which proved its higher accuracy in predicting the abrasion loss of manufactured sand concrete. The machine learning models offer some guidance for predicting and enhancing the abrasion resistance of manufactured sand concrete in practical engineering.

Data-driven approach to optimizing property management strategies: spatial modeling analytics of short-term rentals

Purpose This study aims to explore the impact of locational and seasonal factors on the financial performance of short-term rental properties in Margaret River, Western Australia. It seeks to address the gap in understanding how these factors influence key financial metrics such as average daily rate (ADR) and occupancy rates, providing insights for property managers, investors and policymakers. Design/methodology/approach The research uses a mixed-method approach, integrating advanced predictive modeling techniques, such as Random Forests and Gradient Boosting, with spatial clustering algorithms like density-based spatial clustering of applications with noise (DBSCAN) and ordering points to identify the clustering structure (OPTICS). The study analyzes a comprehensive data set of short-term rental properties between 2012 and 2019. It focuses on locational attributes, seasonal variations and financial outcomes. Findings The findings reveal that properties located near tourist attractions and amenities consistently achieve higher ADRs and occupancy rates, confirming the critical role of location in driving rental demand. Seasonal analysis indicates significant fluctuations in both ADR and occupancy rates, with peaks during high tourist seasons and troughs in off-peak periods. The study underscores the importance of dynamic pricing strategies to optimize revenue and sustain occupancy across different seasons. In addition, it highlights the influence of property features, such as the number of bedrooms and bathrooms, on ADR, while noting that larger properties do not necessarily achieve higher occupancy rates. Research limitations/implications Future research could expand the scope to include different locations and explore the long-term impacts of locational and seasonal factors on property performance. Originality/value This research contributes to the literature by integrating spatial analysis with advanced predictive modeling techniques to provide a nuanced understanding of how locational and seasonal factors impact financial performance in the short-term rental market. It offers a novel application of data analytics within the context of tourism and hospitality management, bridging theoretical frameworks with practical insights.