Gradient Boosting Research Articles

A double-layered fully automated insomnia identification model employing synthetic data generation using MCSA and CTGAN with single-channel EEG signals.

Insomnia was diagnosed by analyzing sleep stages obtained during polysomnography (PSG) recording. The state-of-the-art insomnia detection models that used physiological signals in PSG were successful in classification. However, the sleep stages of unbalanced data in small-time intervals were fed for classification in previous studies. This can be avoided by analyzing the insomnia detection structure in different frequency bands with artificially generated data from the existing one at the preprocessing and post-processing stages. Hence, the paper proposes a double-layered augmentation model using Modified Conventional Signal Augmentation (MCSA) and a Conditional Tabular Generative Adversarial Network (CTGAN) to generate synthetic signals from raw EEG and synthetic data from extracted features, respectively, in creating training data. The presented work is independent of sleep stage scoring and provides double-layered data protection with the utility of augmentation methods. It is ideally suited for real-time detection using a single-channel EEG provides better mobility and comfort while recording. The work analyzes each augmentation layer's performance individually, and better accuracy was observed when merging both. It also evaluates the augmentation performance in various frequency bands, which are decomposed using discrete wavelet transform, and observed that the alpha band contributes more to detection. The classification is performed using Decision Tree (DT), Ensembled Bagged Decision Tree (EBDT), Gradient Boosting (GB), Random Forest (RF), and Stacking classifier (SC), attaining the highest classification accuracy of 94% using RF with a greater Area Under Curve (AUC) value of 0.97 compared to the existing works and is best suited for small datasets.

Analyzing Supervised Machine Learning Models for Classifying Astronomical Objects Using Gaia DR3 Spectral Features

In this paper, we present an analysis of the effectiveness of various machine learning algorithms in classifying astronomical objects using data from the third release (DR3) of the Gaia space mission. The dataset used includes spectral information from the satellite’s red and blue spectrophotometers. The primary goal is to achieve reliable classification with high confidence for symbiotic stars, planetary nebulae, and red giants. Symbiotic stars are binary systems formed by a high-temperature star (a white dwarf in most cases) and an evolved star (Mira type or red giant star); their spectra varies between the typical for these objects (depending on the orbital phase of the object) and present emission lines similar to those observed in PN spectra, which is the reason for this first selection. Several classification algorithms are evaluated, including Random Forest (RF), Support Vector Machine (SVM), Artificial Neural Networks (ANN), Gradient Boosting (GB), and Naive Bayes classifier. The evaluation is based on different metrics such as Precision, Recall, F1-Score, and the Kappa index. The study confirms the effectiveness of classifying the mentioned stars using only their spectral information. The models trained with Artificial Neural Networks and Random Forest demonstrated superior performance, surpassing an accuracy rate of 94.67%.

Prediction of flow stress of Ta–W alloys using machine learning

Abstract The primary aim of this article was to predict the flow stress of Ta–W alloys using the eXtreme Gradient Boosting (XGBoost) machine learning model and to explain the outcome using SHapley Additive exPlanations (SHAP). The article details the effect of temperature, strain rate, and alloying content on the deformation behavior. Though grain size, dislocation density, texture and impurities are also important factors affecting the deformation behavior, these have not been considered in this work. Data and constitutive models from the literature were used to find and compare the predictiveness of the flow stress in Ta–W alloys. XGBoost predicted flow stress with a root mean square error of 12 MPa during training and 40 MPa during testing, while constitutive models such as Johnson–Cook (JC), Zerilli–Armstrong (ZA) and mechanical threshold stress (MTS) models showed a root mean square error of 208, 131 and 149 MPa respectively. The linear correlation between the predicted and experimental flow stress at 10% strain was calculated using the Pearson correlation coefficient and found to be 0.64, 0.93, and 0.70 for JC, ZA and MTS models respectively, while XGBoost showed 0.99 during training and 0.98 during testing. The optimized XGBoost model was validated using five-fold and leave-one-group-out cross-validations. The flow stress at 10% strain was predicted using XGBoost at various temperatures, strain rates, and alloying content. The flow stress was low at temperatures above 1000 K and strain rates below 10−2 s−1. From SHAP analysis, it was found that the base flow stress value (at which the SHAP value is zero) was 477 MPa. For temperatures less than 275 K, strain rates greater than 1 s−1, and alloying content greater than 2.5 wt.% W, the flow stress showed an increase from its base value.

Hybrid LSTM-GBRT based machine learning technique implementation for electric vehicle sales prediction analysis

ABSTRACT The investigation of machine learning models in Europe and India for predicting future sales of electric cars (EVs) worldwide has been the focus of present research. In this study, Long Short-Term Memory (LSTM), Support Vector Regression (SVR), Autoregressive Integrated Moving Average (ARIMA), Linear Regression, hybrid long- and short-term memory neural network model and gradient boosted regression trees model (LSTM-GBRT) models based machine learning techniques are compared for performance analysis to predict EV sales. Moreover, in this study, a hybrid LSTM-GBRT technique was developed, which forecasts EV sales, particularly for Battery Electric Vehicles (BEVs) and Plug-in Hybrid Electric Vehicles (PHEVs) over the next five years. During the prediction of EV sales, it has been validated that hybrid LSTM-GBRT technique has better predictive power. In addition, the present study attempts to provide estimated sales statistics for the designated locations. The results indicate that EV industry is evolving, with India showing a strong growth rate in BEV sales and establishing itself as a major participant in the global EV scenario by showing a notable increase in its share of worldwide EV sales volume by 2027. Conversely, the performance of EV sales in Europe has been analysed and validated through the proposed LSTM-GBRT model.

Feature selection with annealing for forecasting financial time series

Stock market and cryptocurrency forecasting is very important to investors as they aspire to achieve even the slightest improvement to their buy-or-hold strategies so that they may increase profitability. However, obtaining accurate and reliable predictions is challenging, noting that accuracy does not equate to reliability, especially when financial time-series forecasting is applied owing to its complex and chaotic tendencies. To mitigate this complexity, this study provides a comprehensive method for forecasting financial time series based on tactical input–output feature mapping techniques using machine learning (ML) models. During the prediction process, selecting the relevant indicators is vital to obtaining the desired results. In the financial field, limited attention has been paid to this problem with ML solutions. We investigate the use of feature selection with annealing (FSA) for the first time in this field, and we apply the least absolute shrinkage and selection operator (Lasso) method to select the features from more than 1000 candidates obtained from 26 technical classifiers with different periods and lags. Boruta (BOR) feature selection, a wrapper method, is used as a baseline for comparison. Logistic regression (LR), extreme gradient boosting (XGBoost), and long short-term memory are then applied to the selected features for forecasting purposes using 10 different financial datasets containing cryptocurrencies and stocks. The dependent variables consisted of daily logarithmic returns and trends. The mean-squared error for regression, area under the receiver operating characteristic curve, and classification accuracy were used to evaluate model performance, and the statistical significance of the forecasting results was tested using paired t-tests. Experiments indicate that the FSA algorithm increased the performance of ML models, regardless of problem type. The FSA hybrid models showed better performance and outperformed the other BOR models on seven of the 10 datasets for regression and classification. FSA-based models also outperformed Lasso-based models on six of the 10 datasets for regression and four of the 10 datasets for classification. None of the hybrid BOR models outperformed the hybrid FSA models. Lasso-based models, excluding the LR type, were comparable to the best models for six of the 10 datasets for classification. Detailed experimental analysis indicates that the proposed methodology can forecast returns and their movements efficiently and accurately, providing the field with a useful tool for investors.

Machine learning-based analytical approach for mechanical analysis of composite hydrogen storage tanks under internal pressure

This study is a practical exploration of the application of machine learning for the mechanical analysis of filament-wound thin-composite hydrogen storage tanks under internal pressure. Our innovative approach seamlessly integrates classical laminate theory, comprehensive parametric analysis, and machine learning to advance the state-of-the-art in composite tank design. This versatile framework holds promise for addressing challenges in other structurally complex systems. A comprehensive parametric study was then conducted to explore the influence of key design parameters, such as internal pressure, tank radius, order of layer, and layer orientation. Pearson's correlation analysis was used to determine the parameters that had the most significant impact on the mechanical behavior of the tank. In addition, five machine learning models, namely, Extreme Gradient Boosting, Adaptive Boosting, Artificial Neural Network, Support Vector Machine, and Random Forest, were examined to determine their predictive capability for the mechanical performance of these composite tanks. Key findings demonstrate XGBoost's superior predictive capabilities, achieving a remarkable 99% accuracy in predicting stress in the x-direction compared to Random Forest's 96%. XGBoost also outperformed Random Forest in terms of Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE) with values of 26 and 0.02, respectively, versus 207 and 0.16. These results underscore XGBoost's potential to revolutionize hydrogen storage tank design. The development of a user-friendly GUI application further enhances the practicality of this research, allowing engineers and composite designers to efficiently simulate the mechanical behavior of these tanks in real time by adjusting the cursors for variables such as the pressure, radius, layer order, and layer orientation. This tool, with its intuitive interface and elimination of manual calculations, promises significant enhancements in the design process, providing a platform for the rapid assessment and optimization of tank architectures, thereby improving the efficiency and reliability of the design process.

A data science approach to optimize ADHD assessment with the BRIEF-2 questionnaire.

Attention deficit hyperactivity disorder (ADHD) is a prevalent neurodevelopmental disorder. A key challenge associated with this condition is achieving an early diagnosis. The current study seeks to anticipate and delineate the assessments offered by both parents and teachers concerning a child's behavior and overall functioning with the Behavior Rating Inventory of Executive Function-2 (BRIEF-2). Mothers, fathers, and teachers of 59 children diagnosed or in the process of being assessed for ADHD participated in this study. The responses provided by 59 mothers, 59 fathers, and 57 teachers to the BRIEF-2 questionnaire were collected. The performance of various feature selection techniques, including Lasso, decision trees, random forest, extreme gradient boosting, and forward stepwise regression, was evaluated. The results indicate that Lasso stands out as the optimal method for our dataset, striking an ideal balance between accuracy and interpretability. A repeated validation analysis reveals an average positive correlation exceeding 0.5 between the inattention/hyperactivity scores reported by informants (mother, father, or teacher) and the predictions derived from Lasso. This performance is achieved using only approximately 18% of the BRIEF-2 items. These findings underscore the usefulness of variable selection techniques in accurately characterizing a patient's condition while employing a small subset of assessment items. This efficiency is particularly valuable in time-constrained settings and contributes to improving the comprehension of ADHD.

Dual-Stage Analysis Combining Structural Equation Modeling and Machine Learning for Low-Carbon Travel Intention of Urban Residents

Guiding urban residents to travel with low carbon is an important measure to reduce carbon emissions and promote sustainable urban development. To explore the important factors affecting urban residents’ low-carbon travel intention, this study proposes a composite model of urban residents’ low-carbon travel intention based on the composite framework of the theory of planned behavior and value belief norm theory. A total of 398 valid pieces of data were collected in Qingdao, China. Structural equation modeling (SEM) was applied to empirically analyze the data to identify the predictors that have a significant effect on low-carbon travel intention. Then, two machine learning methods, artificial neural networks (ANN) and extreme gradient boosting (XGBoost), were used to conduct sensitivity analysis on the SEM results to identify the determinants that affect residents’ low-carbon travel intention. The results showed that personal norm, attitude, subjective norm, and perceived behavioral control have a significant direct effect on residents’ low-carbon travel intention. Policy factors can indirectly affect low-carbon travel intention, through mediating variables. Environmental awareness and travel time perception have a direct effect on both residents’ attitude and subjective norms. In addition, the explained variance (R2) of low-carbon travel intention by ANN and XGBoost is 0.75 and 0.77, respectively. The Root Square Mean Error (RMSE) in both models are small, which verifies the effectiveness of the two machine learning methods. The results can provide a reference basis for policymakers to prompt urban sustainable development.

Advanced Landslide Detection Using Machine Learning and Remote Sensing Data

Landslides can cause severe damage to infrastructure and human life, making early detection and warning systems critical for mitigating their impact. In this study, we propose a machine learning approach for landslide detection using remote sensing data and topographical features. We evaluate the performance of several machine learning algorithms, including Tree, Random Forest, Gradient Boosting, Logistic Regression, Naïve Bayes, AdaBoost, Neural Network, SGD, kNN, and SVM, on a dataset of remote sensing images and topographical features from the Sikkim region in Malaysia. The results show that the SVM algorithm outperforms the other algorithms with an accuracy of 96.7% and a F1 score of 0.97. The study demonstrates the potential of machine learning algorithms for landslide detection, which can help improve early warning systems and reduce the impact of landslides.

Oxygen Content Control in the Electroslag Remelting Process: An Incremental Learning Strategy Based on Optimized Wasserstein Generative Adversarial Network with Gradient Penalty Data Augmentation

Electroslag remelting (ESR) is essential for producing high‐end special steel, but its complex process and numerous influencing factors make quality control challenging. This study addresses oxygen content control during ESR using a big data machine learning approach. An incremental learning strategy is proposed based on an optimized Wasserstein generative adversarial network with gradient penalty (WGAN‐GP) for data enhancement, focusing on G20Cr2Ni4A bearing steel. The WGAN‐GP model enhances time‐series data and metadata, utilizing long short‐term memory networks, fully connected networks, and attention mechanisms. The effectiveness of data enhancement is verified using a deep neural network classifier and statistical methods. Data is divided into historical and data streams, with an incremental learning strategy based on histogram gradient boosting regression trees to prevent catastrophic forgetting and improve efficiency through knowledge distillation and real‐time hyperparameter adjustment. Results show that the data augmentation method significantly improves model generalization and accuracy in small sample metallurgy. The incremental learning strategy enhances prediction accuracy for oxygen content, contributing to better cleanliness quality of electroslag steel. This study offers a novel approach for addressing small sample challenges in metallurgical processes.

BioPred: An R package for biomarkers analysis in precision medicine.

The R package BioPred offers a suite of tools for subgroup and biomarker analysis in precision medicine. Leveraging Extreme Gradient Boosting (XGBoost) along with propensity score weighting and A-learning methods, BioPred facilitates the optimization of individualized treatment rules (ITR) to streamline subgroup identification. BioPred also enables the identification of predictive biomarkers and obtaining their importance rankings. Moreover, the package provides graphical plots tailored for biomarker analysis. This tool enables clinical researchers seeking to enhance their understanding of biomarkers and patient population in drug development. The package is available at CRAN and https://github.com/deeplearner0731/BioPred. Supplementary data are available at Bioinformatics online.

Genotype-specific responses to in vitro drought stress in myrtle (Myrtus communis L.): integrating machine learning techniques.

Myrtle (Myrtus communis L.), native to the Mediterranean region of Türkiye, is a valuable plant with applications in traditional medicine, pharmaceuticals, and culinary practices. Understanding how myrtle responds to water stress is essential for sustainable cultivation as climate change exacerbates drought conditions. This study investigated the performance of selected myrtle genotypes under in vitro drought stress by employing tissue culture techniques, rooting trials, and acclimatization processes. Genotypes were tested under varying polyethylene glycol (PEG) concentrations (1%, 2%, 4%, and 6%). Machine learning (ML) algorithms, including Gaussian process (GP), support vector machine (SVM), Random Forest (RF), and Extreme Gradient Boosting (XGBoost), were utilized to model and predict micropropagation and rooting efficiency. The research revealed a genotype-dependent response to drought stress. Black-fruited genotypes exhibited higher micropropagation rates compared to white-fruited ones under stress conditions. The application of ML models successfully predicted micropropagation and rooting efficiency, providing insights into genotype performance. The findings suggest that selecting drought-tolerant genotypes is crucial for enhancing myrtle cultivation. The results underscore the importance of genotype selection and optimization of cultivation practices to address climate change impacts. Future research should explore the molecular mechanisms of stress responses to refine breeding strategies and improve resilience in myrtle and similar economically important crops.

A STUDY OF THE STANDARDIZED PRECIPITATION INDEX (SPI) AND MACHINE LEARNING TECHNIQUES FOR DROUGHT PREDICTION IN THE STATE OF PARAÍBA, BRAZIL

This study aimed to identify and analyze droughts in Paraíba, using the Standardized Precipitation Index (SPI) and machine learning algorithms for predicting SPI for the subsequent years (2020-2021) at six rainfall stations distributed across the mesoregions of Paraíba. The Precipitation data were downloaded from the Global Precipitation Climatology Centre (GPCC) and the National Oceanic and Atmospheric Administration (NOAA) database, covering the period from 1991 to 2019. Three machine learning algorithms were selected based on their ability to fit historical SPI data: Extra Trees Regressor, Gradient Boosting Regressor, and Random Forest Regressor. The applied machine learning models yielded satisfactory results, with the Extra Trees Regressor consistently producing the highest R² value across all stations, indicating high data explainability. The predictions were analyzed to determine their accuracy and reliability, providing valuable insights into precipitation variability and drought occurrence in different mesoregions of Paraíba. In conclusion, this study contributed to understanding climate variability and its implications in Paraíba, offering valuable insights into drought occurrence and the importance of adaptive approaches to mitigate adverse impacts. The application of SPI and machine learning techniques proved effective in analyzing and predicting precipitation, providing an objective approach to characterizing drought and rainfall intensity in specific regions.

Modeling motorcycle crash-injury severity utilizing explainable data-driven approaches

ABSTRACT Motorcycle crashes remain a significant public safety concern, requiring diverse analytical approaches to inform countermeasures. This study uses machine learning to analyze injury severity in crashes in Thailand from 2018 to 2020. Traditional and advanced models, including including random forest (RF), support vector machine (SVM), deep neural network (DNN), recurrent neural network (RNN), long short-term memory (LSTM), and eXtreme gradient boosting (XGBoost), were compared. Hyperparameter tuning via GridSearchCV optimized performance. XGBoost, with a tradeoff score of 105.65%, outperformed other models in predicting severe and fatal injuries. SHapley Additive exPlanations (SHAPs) identified significant risk factors including speeding, drunk driving, two-lane roads, unlit conditions, head-on and truck collisions, and nighttime crashes. Conversely, factors such as barrier medians, flashing traffic signals, sideswipes, rear-end crashes, and wet roads were associated with reduced severity. These findings suggest opportunities for integrated infrastructure improvements and expanded rider training and education programs to address behavioral risks.

Prediction of Ductile Damage in Composite Material Used in Type IV Hydrogen Tanks by Artificial Neural Network and Machine Learning with Finite Element Modeling Approach

This study investigates the degradation process of composite materials used in high‐pressure hydrogen storage vessels by employing advanced computational techniques. A recurrent neural network, specifically a bidirectional long short‐term memory (Bi‐LSTM) network, is utilized to predict the temporal evolution of ductile damage. The key degradation features are extracted from finite element modeling (FEM) computations using group method of data handling algorithms and treated as time‐series data. Results demonstrate that the Bi‐LSTM network can accurately undergo both elastic and plastic behaviors of the composite under tensile strength. Additionally, traditional machine learning (ML) algorithms such as extreme gradient boosting and random forest are employed to forecast strain degradation, showing promising results. This hybrid approach combining FEM, ML, and deep learning provides a comprehensive method for predicting the degradation of composite materials, offering significant potential for optimizing the design and durability of hydrogen storage vessels.

Image-based assessment of seismic damage in RC exterior beam-column joints

Seismic damage to reinforced concrete (RC) beam-column joints in reinforced concrete (RC) structures significantly impacts their stability and safety after an earthquake. Cracks caused by earthquakes weaken these joints, reducing their ductility, strength, and stiffness. Therefore, a systematic method for assessing damage in RC beam-column joints is crucial. This study introduces a new method that uses image analysis, along with ranking algorithms and machine learning, to assess seismic damage in RC exterior beam-column joints. The study identified three key indicators of severe damage in joints: drift ratio, strength, and stiffness. The method uses image analysis to extract features from preprocessed, binary images of cracks. These features capture both the texture and geometry of the cracks. The chosen features, including crack area, aspect ratio, and distribution of line widths, along with structural parameters like geometry and bond index, were selected using F-test and MRMR algorithms. Seismic damage assessments were conducted on 115 images from 37 specimens using four machine learning models (Regression Trees, Support Vector Machine, Gaussian Process Regression, and Gradient Boosting). The study achieved a high R-squared value of 0.80, indicating strong accuracy, for assessing seismic damage based on drift ratio using the image-based method. This demonstrates that image analysis techniques can effectively link surface crack patterns to quantifiable image features, leading to a more precise assessment of seismic damage.

An explainable hybrid framework for estimating daily reference evapotranspiration: Combining extreme gradient boosting with Nelder-Mead method

Accurate estimation of reference evapotranspiration (ETo) is essential for effective water resources management, irrigation system design, and various hydrological and agricultural applications. This study employed extreme gradient boosting (XGBoost) model, signal decomposition techniques, and XGBoost coupled with Nelder–Mead (NM) method to enhance ETo prediction across two meteorological stations in Iran. This study proposed a novel framework which lies in its comprehensive integration of advanced techniques to create a model that is both interpretable and highly accurate for ETo estimation. For this aim, sixty meteorological variables were categorized into solar and cloud-based, temperature-based, wind and humidity-based, and pressure-based groups to analyze their effect on ETo estimation. Feature selection methods, including the Gradient Boosting Machine, Kendall’s Tau, and Relief Algorithm, were employed to identify the most influential predictors. Variables with normalized weights equal to or greater than 0.9 were selected for model input, resulting in the top 10% of variables being utilized. The XGBoost models were then developed using these selected inputs (level 1), a wavelet-based hybrid was developed according to the most effective features (level 2), and the XGBoost model was coupled by NM algorithm (level 3) for ETo estimation. Results of estimated ETo by levels 1 to 3 were compared with four common empirical approaches. The results indicated that Kendall’s Tau-based feature selection provided the most accurate predictions in Shiraz, achieving an RMSE of 0.721 (mm/day) for solar and cloud-based variables during the test phase. Additionally, the application of wavelet analysis further refined the model inputs, which enhanced ETo estimation in most variable groups. Integrating the NM algorithm with XGBoost demonstrated significant improvements in ETo estimation, where it could improve RMSE to 0.091 and 0.155 (mm/day) for testing section in Fasa and Shiraz, respectively.

Improving port state control through a transfer learning-enhanced XGBoost model

With the advancements in modern information technology, Port State Control (PSC) inspections, as a crucial measure to protect ship safety and the marine environment, are undergoing an intelligent transformation. This paper aims to streamline the selection process for inspecting high-risk ships by employing a data-driven model to predict the number of deficiencies in ships arriving at ports. A transfer learning-enhanced eXtreme Gradient Boosting (XGBoost) model is proposed by innovatively incorporating sample similarity calculations to adapt the model to the unique characteristics of the target port. This novel model enables the integration of relevant data from other ports, enhancing the predictive performance of the model to specific port conditions. Utilizing PSC inspection records from ports within the Tokyo Memorandum of Understanding (MoU) and choosing the port of Singapore as the target, numerical experiments demonstrate that the proposed model achieves improvements of 1.81 %, 6.08 %, and 3.60 % in the mean absolute error, mean squared error and root mean squared error, respectively, compared to the standard XGBoost model. Furthermore, across various sizes of training samples, the proposed model outperforms other machine learning models. This work may service as a significant step towards exploring the potential of developing data-driven models based on comprehensive data to assess the risk level of foreign ships arriving at ports, ameliorating the PSC inspection process by aiding PSC officers in identifying substandard ships more effectively.

A retinal detachment based strabismus detection through FEDCNN

Ocular strabismus, a common condition in the present generation is an absolute risk factor for amblyopia and blinding premorbid visual loss. Despite the availability of new optometry tools with eye-tracking data, the issues persist in attaining accuracy and reliability in diagnosing strabismus. These two concerns are specifically accommodated in this study by the proposed novel approach that involves CNNs with eye-tracking datasets from subjects. The presented work aims to improve the accuracy of diagnostics in ophthalmology utilizing the integration of the further proposed algorithms into an automatic strabismus detection system. For this purpose, the proposed FedCNN model combines the CNN with eXtreme Gradient Boosting (XGBoost) and uses the Gaze deviation (GaDe) images to capture dynamic eye movements. This method tries to make the feature extraction as accurate as possible in its best working state to enhance the diagnosis precision. The model proves to be accurate, reaching 95.2%, which is even more prominent because of the more or less detailed connection layer of the CNN, which is used for the selection of features designated for such tasks of strabismus recognition. The presented method has the potential of shifting the approach to diagnosing diseases of the eyes in more or less half of the patients.

Interpretable machine learning for predicting heavy metal removal efficiency in electrokinetic soil remediation

Electrokinetic remediation (EKR) presents a promising approach for polluted soil remediation, leveraging electric fields to mobilize contaminants and facilitate their removal. Accurate efficiency prediction is crucial for optimizing EKR processes, reducing costs, and minimizing environmental impact. However, the EKR efficiency is influenced by numerous complex factors, making it challenging to predict outcomes reliably. This study introduces a novel approach for developing interpretable machine learning (ML) models tailored for efficient and unbiased EKR efficiency prediction based on material properties and experimental conditions. A large experimental dataset comprising 185 tests was compiled, encompassing various EKR parameters, heavy metal concentration, soil characteristics. A structured ML workflow was devised, evaluating diverse ML algorithms and employing model-agnostic interpretation methodologies to uncover intrinsic correlations between removal efficiency and pertinent attributes. Among the investigated ML models, Extreme Gradient Boosting (XGB) exhibited the best performance. Bayesian optimization was applied to further enhance its capabilities. The optimized XGB model demonstrated superior predictive accuracy on the test set, with an RMSE of 0.255, an MAE of 0.158, and an R² of 0.82. Moreover, SHapley Additive exPlanations analysis was employed to interpret the contributions of the input variables. Among inputs, electrode distance emerged as the most significant factor influencing EKR efficiency, followed by area, electrolyte species, and remediation duration. This research not only advances the predictive capabilities for EKR efficiency but also provides crucial insights into the underlying factors influencing remediation success, paving the way for more efficient and scalable applications in environmental management.