Accurate reservoir permeability determination is crucial in hydrocarbon exploration and production.Conventional methods relying on empirical correlations and assumptions often result in high costs,time consumption,ina...Accurate reservoir permeability determination is crucial in hydrocarbon exploration and production.Conventional methods relying on empirical correlations and assumptions often result in high costs,time consumption,inaccuracies,and uncertainties.This study introduces a novel hybrid machine learning approach to predict the permeability of the Wangkwar formation in the Gunya oilfield,Northwestern Uganda.The group method of data handling with differential evolution(GMDH-DE)algorithm was used to predict permeability due to its capability to manage complex,nonlinear relationships between variables,reduced computation time,and parameter optimization through evolutionary algorithms.Using 1953 samples from Gunya-1 and Gunya-2 wells for training and 1563 samples from Gunya-3 for testing,the GMDH-DE outperformed the group method of data handling(GMDH)and random forest(RF)in predicting permeability with higher accuracy and lower computation time.The GMDH-DE achieved an R^(2)of 0.9985,RMSE of 3.157,MAE of 2.366,and ME of 0.001 during training,and for testing,the ME,MAE,RMSE,and R^(2)were 1.3508,12.503,21.3898,and 0.9534,respectively.Additionally,the GMDH-DE demonstrated a 41%reduction in processing time compared to GMDH and RF.The model was also used to predict the permeability of the Mita Gamma well in the Mandawa basin,Tanzania,which lacks core data.Shapley additive explanations(SHAP)analysis identified thermal neutron porosity(TNPH),effective porosity(PHIE),and spectral gamma-ray(SGR)as the most critical parameters in permeability prediction.Therefore,the GMDH-DE model offers a novel,efficient,and accurate approach for fast permeability prediction,enhancing hydrocarbon exploration and production.展开更多
The methods of network attacks have become increasingly sophisticated,rendering traditional cybersecurity defense mechanisms insufficient to address novel and complex threats effectively.In recent years,artificial int...The methods of network attacks have become increasingly sophisticated,rendering traditional cybersecurity defense mechanisms insufficient to address novel and complex threats effectively.In recent years,artificial intelligence has achieved significant progress in the field of network security.However,many challenges and issues remain,particularly regarding the interpretability of deep learning and ensemble learning algorithms.To address the challenge of enhancing the interpretability of network attack prediction models,this paper proposes a method that combines Light Gradient Boosting Machine(LGBM)and SHapley Additive exPlanations(SHAP).LGBM is employed to model anomalous fluctuations in various network indicators,enabling the rapid and accurate identification and prediction of potential network attack types,thereby facilitating the implementation of timely defense measures,the model achieved an accuracy of 0.977,precision of 0.985,recall of 0.975,and an F1 score of 0.979,demonstrating better performance compared to other models in the domain of network attack prediction.SHAP is utilized to analyze the black-box decision-making process of the model,providing interpretability by quantifying the contribution of each feature to the prediction results and elucidating the relationships between features.The experimental results demonstrate that the network attack predictionmodel based on LGBM exhibits superior accuracy and outstanding predictive capabilities.Moreover,the SHAP-based interpretability analysis significantly improves the model’s transparency and interpretability.展开更多
Accurate prediction of shield tunneling-induced settlement is a complex problem that requires consideration of many influential parameters.Recent studies reveal that machine learning(ML)algorithms can predict the sett...Accurate prediction of shield tunneling-induced settlement is a complex problem that requires consideration of many influential parameters.Recent studies reveal that machine learning(ML)algorithms can predict the settlement caused by tunneling.However,well-performing ML models are usually less interpretable.Irrelevant input features decrease the performance and interpretability of an ML model.Nonetheless,feature selection,a critical step in the ML pipeline,is usually ignored in most studies that focused on predicting tunneling-induced settlement.This study applies four techniques,i.e.Pearson correlation method,sequential forward selection(SFS),sequential backward selection(SBS)and Boruta algorithm,to investigate the effect of feature selection on the model’s performance when predicting the tunneling-induced maximum surface settlement(S_(max)).The data set used in this study was compiled from two metro tunnel projects excavated in Hangzhou,China using earth pressure balance(EPB)shields and consists of 14 input features and a single output(i.e.S_(max)).The ML model that is trained on features selected from the Boruta algorithm demonstrates the best performance in both the training and testing phases.The relevant features chosen from the Boruta algorithm further indicate that tunneling-induced settlement is affected by parameters related to tunnel geometry,geological conditions and shield operation.The recently proposed Shapley additive explanations(SHAP)method explores how the input features contribute to the output of a complex ML model.It is observed that the larger settlements are induced during shield tunneling in silty clay.Moreover,the SHAP analysis reveals that the low magnitudes of face pressure at the top of the shield increase the model’s output。展开更多
Accurately revealing the spatial heterogeneity in the trade-offs and synergies of land use functions(LUFs)and their driving factors is imperative for advancing sustainable land utilization and optimizing land use plan...Accurately revealing the spatial heterogeneity in the trade-offs and synergies of land use functions(LUFs)and their driving factors is imperative for advancing sustainable land utilization and optimizing land use planning.This is especially critical for ecologically vulnerable inland river basins in arid regions.However,existing methods struggle to effectively capture complex nonlinear interactions among environmental factors and their multifaceted relationships with trade-offs and synergies of LUFs,especially for the inland river basins in arid regions.Consequently,this study focused on the middle reaches of the Heihe River Basin(MHRB),an arid inland river basin in northwestern China.Using land use,socioeconomic,meteorological,and hydrological data from 2000 to 2020,we analyzed the spatiotemporal patterns of LUFs and their trade-off and synergy relationships from the perspective of production,living,ecological functions.Additionally,we employed an integrated Extreme Gradient Boosting(XGBoost)-SHapley Additive exPlanations(SHAP)framework to investigate the environmental factors influencing the spatial heterogeneity in the trade-offs and synergies of LUFs.Our findings reveal that from 2000 to 2020,the production,living,and ecological functions of land use within the MHRB exhibited an increasing trend,demonstrating a distinct spatial pattern of''high in the southwest and low in the northeast''.Significant spatial heterogeneity defined the trade-off and synergistic relationships,with trade-offs dominating human activity-intensive oasis areas,while synergies prevailed in other areas.During the study period,synergistic relationships between production and living functions and between production and ecological functions were relatively robust,whereas synergies in living-ecological functions remained weaker.Natural factors(digital elevation model(DEM),annual mean temperature,Normalized Difference Vegetation Index(NDVI),and annual precipitation)emerged as the primary factors driving the trade-offs and synergies of LUFs,followed by socioeconomic factors(population density,Gross Domestic Product(GDP),and land use intensity),while distance factors(distance to water bodies,distance to residential areas,and distance to roads)exerted minimal influence.Notably,the interactions among NDVI,annual mean temperature,DEM,and land use intensity exerted the most substantial impacts on the relationships among LUFs.This study provides novel perspectives and methodologies for unraveling the mechanisms underlying the spatial heterogeneity in the trade-offs and synergies of LUFs,offering scientific insights to inform regional land use planning and sustainable natural resource management in inland river basins in arid regions.展开更多
Ensuring independent mobility for older adults has become a public health and social concern in China owing to its rapidly aging population.To explore independent mobility trends among older adults and the impact of s...Ensuring independent mobility for older adults has become a public health and social concern in China owing to its rapidly aging population.To explore independent mobility trends among older adults and the impact of sociodemo-graphic characteristics in recent years,this study used data from the Chinese Longitudinal Healthy Longevity Survey from 2012 to 2018,combined with binomial logit regression and CatBoost-Shapley additive explanation(SHAP)method to analyze the relationship between independent mobility and sociodemographic characteristics under bus and walking-oriented environments.Study findings indicated that age and gender significantly affected the independent mobility of older adults.Policymaking should prioritize the needs of older adults,focusing on age and gender differ-ences.Additionally,living expense adequacy significantly influenced independent mobility.Policies should substan-tially support economically disadvantaged older adults,en-suring their basic needs are met through subsidies and other measures.Moreover,the study found a notable impact of widowhood on independent mobility,suggesting enhanced social care and mental health support for widowed older adults,especially those who are long-lived.The outcomes of this study provided evidence for policymakers,which are beneficial for developing elderly-friendly travel policies to ensure and enhance the quality of life and independent mo-bility of older adults.展开更多
In the arid regions of Northwest China,vegetation cover plays a crucial role in maintaining unique terrestrial ecosystems.Vegetation growth is highly sensitive to variations in topographical factors,and the influence ...In the arid regions of Northwest China,vegetation cover plays a crucial role in maintaining unique terrestrial ecosystems.Vegetation growth is highly sensitive to variations in topographical factors,and the influence of topography on vegetation cover has attracted increasing attention.This study analyzed vegetation dynamics and their relationship with topography in the Tianshan Mountains of China using Landsat Normalized Difference Vegetation Index(NDVI)data during 2000–2022 and Shuttle Radar Topography Mission(SRTM)-derived topographical factors(elevation,slope,and aspect).Theil-Sen slope estimation and Mann-Kendall trend tests were applied to quantify temporal changes in vegetation,while a terrain area correction coefficient(K)was used to assess spatial associations of vegetation with topography.Random Forest(RF)regression and SHapley Additive exPlanations(SHAP)analysis evaluated the relative importance of topographical factors in shaping vegetation cover(multi-year mean NDVI)distribution.Key findings included that over the 23-a period,59.46%of the vegetated area exhibited significant improvement(P<0.05),with the southern Tianshan Mountains showing the most pronounced increase(70.59%),whereas vegetation degradation(3.10%)was primarily concentrated in river valleys with intensive human activities.RF-SHAP analysis revealed that elevation is the primary driver of vegetation cover patterns,explaining 52.00%of the NDVI variation.The peak NDVI(0.42)occurred at elevations between 2800 and 3200 m.Slope and aspect also significantly influenced vegetation distribution,and higher NDVI values and greater improvement trends were observed on shady(north-facing)slopes compared to sunny(south-facing)slopes.K-index analysis indicated pronounced vegetation change—both degradation and improvement—in areas with elevations between 1100 and 2800 m and slopes exceeding 5°,particularly on sunny slopes.Low-elevation desert areas in the southern Tianshan Mountains were highly susceptible to degradation.This study underscores the critical role of topography in regulating vegetation cover and its spatiotemporal dynamics,providing a scientific basis for sustainable management of arid mountain ecosystems.展开更多
Within the context of global climate change and rapid urbanization,increasing urban resilience(UR)is especially important in the arid region of Northwest China(ANC),where fragile ecosystems and an uneven water distrib...Within the context of global climate change and rapid urbanization,increasing urban resilience(UR)is especially important in the arid region of Northwest China(ANC),where fragile ecosystems and an uneven water distribution create significant sustainability challenges.In this study,a coupled UR-water ecosystem services(WESs)framework was developed on the basis of 1-km resolution remote sensing data for the 2000–2020 period obtained from the Landsat series,Defense Meteorological Satellite Program(DMSP)/Operational Linescan System(OLS),and Global Precipitation Measurement(GPM),among other sources.Within the framework,the Integrated Valuation of Ecosystem Services and Tradeoffs(InVEST)model was incorporated to provide a WES indicator system.Moreover,entropy weighting was employed to quantify both UR and WES indicators;the coupling coordination degree(CCD)model was used to measure the coupled relationship between UR and WESs;an extreme gradient boosting(XGBoost)-SHapley Additive exPlanations(SHAP)interpretation approach was adopted to identify key drivers and underlying mechanisms;and Geographically Weighted Regression(GWR)was applied to capture spatial distribution characteristics of major driving factors.The results indicated that UR steadily increased from 4.60×10^(-3) to 10.24×10^(-3),whereas WESs followed an inverted V-shaped trend,with a peak value observed in 2010(11.84×10^(-3)).The CCD remained consistently low(mean:0.0166–0.0246)and exhibited considerable spatial heterogeneity.Notably,the degree of coordination was greater in the oasis and mountain core areas but significantly lower at desert areas.XGBoost-SHAP model analysis revealed six key drivers influencing various states of the CCD between UR and WESs systems.The contributions of these factors could be ranked as follows:water yield(WY;24.30%)>farmland area per capita(FP;21.10%)>gross domestic product(GDP)per capita(GDPC;19.00%)>soil retention(SR;14.90%)>population density(PD;5.42%)>water purification(WP;4.40%).In contrast,in UR system,WP(53.66%)and SR(31.62%)served as the dominant drivers.Moreover,the dominant drivers shifted from a combination of natural and socioeconomic factors in StateⅠ(sustainable high resilience)to primarily socioeconomic factors in StateⅢ(unsustainable low resilience).SR and WP exerted positive moderating effects,whereas socioeconomic factors such as GDPC and PD exerted inhibitory effects on the coordination relationship.This research highlights that UR in the ANC region is limited mainly by water scarcity,weak feedback loops,and spatial variability,emphasizing the need for tailored intervention strategies.展开更多
Arid mountain ecosystems are highly sensitive to hydrothermal stress and land use intensification,yet where net primary productivity(NPP)degradation is likely to persist and what drives it remain unclear in the Tiansh...Arid mountain ecosystems are highly sensitive to hydrothermal stress and land use intensification,yet where net primary productivity(NPP)degradation is likely to persist and what drives it remain unclear in the Tianshan Mountains of Northwest China.We integrated multi-source remote sensing with the Carnegie–Ames–Stanford Approach(CASA)model to estimate NPP during 2000–2020,assessed trend persistence using the Hurst exponent,and identified key drivers and nonlinear thresholds with Extreme Gradient Boosting(XGBoost)and SHapley Additive exPlanations(SHAP).Total NPP averaged 55.74 Tg C/a and ranged from 48.07 to 65.91 Tg C/a from 2000 to 2020,while regional mean NPP rose from 138.97 to 160.69 g C/(m^(2)·a).Land use transfer analysis showed that grassland expanded mainly at the expense of unutilized land and that cropland increased overall.Although NPP increased across 64.11%of the region during 2000–2020,persistence analysis suggested that 53.93%of the Tianshan Mountains was prone to continued NPP decline,including 36.41%with significant projected decline and 17.52%with weak projected decline;these areas formed degradation hotspots concentrated in the central and northern Tianshan Mountains.In contrast,potential improvement was limited(strong persistent improvement:4.97%;strong anti-persistent improvement:0.36%).Driver attribution indicated that land use dominated NPP variability(mean absolute SHAP value=29.54%),followed by precipitation(16.03%)and temperature(11.05%).SHAP dependence analyses showed that precipitation effects stabilized at 300.00–400.00 mm,and temperature exhibited an inverted U-shaped response with a peak near 0.00°C.These findings indicated that persistent degradation risk arose from hydrothermal constraints interacting with land use conversion,highlighting the need for threshold-informed,spatially targeted management to sustain carbon sequestration in arid mountain ecosystems.展开更多
In the northern Tarim River Basin,the Weigan River Basin is a critical endorheic system characterized by extreme aridity,where drought poses a major natural hazard to agricultural production and ecological stability.T...In the northern Tarim River Basin,the Weigan River Basin is a critical endorheic system characterized by extreme aridity,where drought poses a major natural hazard to agricultural production and ecological stability.This study assessed the future evolution of drought under climate change by employing the standardized moisture anomaly index(SZI)on the basis of multi-model the Coupled Model Intercomparison Project Phase 6(CMIP6)simulations under historical conditions(1970–2014)and future scenarios(shared socioeconomic pathway(SSP)1-2.6,SSP2-4.5,SSP3-7.0 and SSP5-8.5 for 2015–2100).The results show that precipitation–evapotranspiration anomalies are projected to first decline but then increase over time,with increased fluctuations and uncertainty under high-emission scenarios(SSP5-8.5).These trends indicate intensifying drought risks and reveal a strong influence of emission pathways on regional water cycling.Temporal analysis of SZI indicates a transition from wetting to drying under lowand medium-emission pathways(SSP1-2.6 and SSP2-4.5),whereas high-emission scenarios are characterized by persistent drying and increased variability.The significant lower-tail dependence(0.271)observed under SSP2-4.5 and SSP5-8.5 suggests that extreme droughts may be subject to nonlinear co-amplification across scenarios.The frequency of moderate and more severe drought events is expected to increase substantially,especially under SSP5-8.5,where drought occurrence is predicted to extend into spring and autumn and become more evenly distributed throughout the year.Spatially,drought duration shows significant positive autocorrelation across all scenarios,with hot spots consistently concentrated in the southern and southeastern regions of the basin.Random forest analysis,interpreted as association-based pattern attribution,indicates that meteorological variables(precipitation and potential evapotranspiration(PET))make the greatest contributions to the hot spot pattern,followed by topography and soil moisture.Among land use categories,farmland generally shows higher drought sensitivity than other land use types,as reflected by its relative contribution patterns across scenarios.The spatial pattern of drought is statistically structured by climatic forcing,surface conditions,and soil moisture status,reflecting their coupled associations with hot spot occurrence.In addition,a drought spatial uncertainty index was constructed from multi-scenario hot spot maps,revealing spatially heterogeneous structural variability throughout the basin.Correlation analysis further highlights strong internal couplings among environmental variables(e.g.,elevation-linked hydroclimatic gradients and grassland–bare soil contrasts).These findings offer a scientific basis for developing region-specific drought monitoring and adaptation strategies under future climate change conditions.展开更多
Color has emerged as a pivotal factor influencing consumer purchasing decisions in the dried herbal medicine market.To address the issue of significant discoloration of Rhubarb(Rheum rhabarbarum L.)during the drying p...Color has emerged as a pivotal factor influencing consumer purchasing decisions in the dried herbal medicine market.To address the issue of significant discoloration of Rhubarb(Rheum rhabarbarum L.)during the drying process,this study investigates the effects of microwave fixation(MF)and hot-air fixation(HAF)pretreatment methods on the drying characteristics and quality of Rhubarb by ultrasonic synergistic vacuum far-infrared drying(U-VFID),with a primary focus on its color attributes.The results indicate that fixation pretreatment significantly enhances both drying efficiency and product quality,particularly in terms of color retention.Compared to unpretreated Rhubarb,the best comprehensive drying effect was achieved with MF treatment at 60℃for 7 min,which resulted in a 441.18%increase in rhein content,a 58.57%reduction in drying time,and a 48.38%decrease in theΔE value.Furthermore,correlation analysis,and the eXtreme Gradient Boosting(XGBoost)algorithm combined with SHapley Additive exPlanations(SHAP)model,revealed that the color of Rhubarb subjected to various fixation pretreatments in conjunction with U-VFID is primarily influenced by sennoside A content,total phenolic content(TPC),and drying time.This study offers a scientific foundation and theoretical insights for optimizing the quality of dried medicinal plant products and introduces innovative approaches for post-harvest industrial pretreatment of rhizomatous medicinal plants.展开更多
本文聚焦我国青海湖流域的水文过程,基于多年气象和水文动态数据,发展了一种融合概念性水文模型FLEX (FluxExchange)和门控循环单元(Gated Recurrent Unit, GRU)的混合模型对流域内最大支流布哈河的逐日径流进行了模拟和预测。在构建混...本文聚焦我国青海湖流域的水文过程,基于多年气象和水文动态数据,发展了一种融合概念性水文模型FLEX (FluxExchange)和门控循环单元(Gated Recurrent Unit, GRU)的混合模型对流域内最大支流布哈河的逐日径流进行了模拟和预测。在构建混合模型中,采用了三种策略提升模拟精度:引入差分进化自适应算法DREAM(zs)反演水文参数优化FLEX模型;采用变分模态分解(VMD)提取径流数据的信息和特征;利用麻雀搜索算法(SSA)优化深度学习GRU的参数。研究将FLEX模型的模拟结果连同气象数据一起作为神经网络的输入,从而构建了FLEX-VMD-SSA-GRU混合模型。同时,探讨了不同的气象输入条件对模拟结果的影响和贡献:基于7个主要气象要素,由少及多设置了14组输入情景模拟。最后通过SHAP对深度学习方法的结果进行分析,揭示了气象变量对径流长期趋势的贡献和重要度。展开更多
This study addresses gaps in aftershock prediction research by proposing an interpretable hybrid machine learning model that leverages multi-source data.The model overcomes challenges related to the selection of influ...This study addresses gaps in aftershock prediction research by proposing an interpretable hybrid machine learning model that leverages multi-source data.The model overcomes challenges related to the selection of influencing factors,model types,prediction result visualization,and decision mechanism interpretability.It integrates mainshock factors,geological features,site characteristics,and terrain conditions using geospatial information system(GIS)technology.By employing the stacking algorithm to optimize and combine XGBoost and LightGBM models,the proposed model significantly improves the prediction performance.Visualization through aftershock hazard mapping offers a robust tool for aftershock warning.The Shapley additive explanations(SHAP)model is used to explain the decision-making process from both global and local perspectives.Results show that,compared to the optimized XGBoost-CMA_ES and LightGBM-CMA_ES hybrid models,the stacking model achieves area under the curve(AUC)increases of 7.71%and 5.72% on the test set,respectively,with a maximum prediction accuracy of 0.9344.The hazard zoning map identifies high-risk areas mainly around fault lines and near the epicenter.As hazard levels rise,the proportion and density of aftershocks in these areas increase.The SHAP model results highlight the distance to fault as the most critical factor.The study integrates local explanations with on-site investigations,effectively visualizing the contributions of different factors to aftershocks.This research provides new tools and methods for enhancing aftershock warning and response.展开更多
Shield attitudes,essentially governed by intricate mechanisms,impact the segment assembly quality and tunnel axis deviation.In data-driven prediction,however,existing methods using the original driving parameters fail...Shield attitudes,essentially governed by intricate mechanisms,impact the segment assembly quality and tunnel axis deviation.In data-driven prediction,however,existing methods using the original driving parameters fail to present convincing performance due to insufficient consideration of complicated interactions among the parameters.Therefore,a multi-dimensional feature synthesizing and screening method is proposed to explore the optimal features that can better reflect the physical mechanism in predicting shield tunneling attitudes.Features embedded with physical knowledge were synthesized from seven dimensions,which were validated by the clustering quality of Shapley Additive Explanations(SHAP)values.Subsequently,a novel index,Expected Impact Index(EII),has been proposed for screening the optimal features reliably.Finally,a Bayesian-optimized deep learning model was established to validate the proposed method in a case study.Results show that the proposed method effectively identifies the optimal parameters for shield attitude prediction,with an average Mean Squared Error(MSE)deduction of 27.3%.The proposed method realized effective assimilation of shield driving data with physical mechanism,providing a valuable reference for shield deviation control.展开更多
The uncertainty of ocean turbulent mixing parameterization comprises a significant challenge in ocean and climate models. A depth-dependent deep learning ocean turbulent mixing parameterization scheme was proposed wit...The uncertainty of ocean turbulent mixing parameterization comprises a significant challenge in ocean and climate models. A depth-dependent deep learning ocean turbulent mixing parameterization scheme was proposed with the hydrological and microstructure observations conducted in summer 2012 in the shelf sea east of Hainan Island, in South China Sea(SCS). The deep neural network model is used and incorporates the Richardson number Ri, the normalized depth D, the horizontal velocity speed U, the shear S^(2), the stratification N^(2), and the density ρ as input parameters. Comparing to the scheme without parameter D and region division, the depth-dependent scheme improves the prediction of the turbulent kinetic energy dissipation rate ε. The correlation coefficient(r) between predicted and observed lgε increases from 0.49 to 0.62, and the root mean square error decreases from 0.56 to 0.48. Comparing to the traditional physics-driven parameterization schemes, such as the G89 and MG03, the data-driven approach achieves higher accuracy and generalization. The SHapley Additive Explanations(SHAP) framework analysis reveals the importance descending order of the input parameters as: ρ, D, U, N^(2), S^(2), and Ri in the whole depth, while D is most important in the upper and bottom boundary layers(D≤0.3&D≥0.65) and least important in middle layer(0.3<D<0.65). The research shows applicability of constructing deep learning-based ocean turbulent mixing parameterization schemes using limited observational data and well-established physical processes.展开更多
Accurate and interpretable prediction of shield tunnel-induced settlement poses a significant challenge due to the complex interplay of various influencing factors.This paper proposes a novel interpretable hybrid mode...Accurate and interpretable prediction of shield tunnel-induced settlement poses a significant challenge due to the complex interplay of various influencing factors.This paper proposes a novel interpretable hybrid model that combines complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN),slime mold algorithm(SMA),and least squares support vector machine(LSSVM)to enhance prediction accuracy and model transparency.The CEEMDAN method,optimized by SMA,decomposes settlement data into intrinsic mode functions(IMFs)and residuals,thereby reducing data noise.The LSSVM,also optimized by SMA,is then applied to predict each IMF and residual.The final settlement prediction is derived from the aggregation of these results.The model was rigorously validated using the Changsha(China)and Singapore Metro projects,demonstrating superior performance to traditional machine learning models.The evaluation metrics,including root mean square error(RMSE),mean absolute error(MAE),mean absolute percentage error(MAPE),and coefficient of determination(R2),underscore the model's effectiveness.The model achieved the lowest error rates and highest accuracy across these metrics.Notably,Shapley additive explanations(SHAP)provided insights into the model's decision-making process,identifying shield stoppage and moisture content as the most influential factors in settlement prediction.This study contributes to the advancement of the methodological framework for predicting tunnel settlement.It addresses the discrepancy between prediction accuracy and interpretability,providing a robust tool for practical engineering applications.展开更多
基金supported by the Major National Science and Technology Programs in the“Thirteenth Five-Year”Plan period(Grant No.2017ZX05032-002-004)the Innovation Team Funding of Natural Science Foundation of Hubei Province,China(Grant No.2021CFA031)the Chinese Scholarship Council(CSC)and Silk Road Institute for their support in terms of stipend.
文摘Accurate reservoir permeability determination is crucial in hydrocarbon exploration and production.Conventional methods relying on empirical correlations and assumptions often result in high costs,time consumption,inaccuracies,and uncertainties.This study introduces a novel hybrid machine learning approach to predict the permeability of the Wangkwar formation in the Gunya oilfield,Northwestern Uganda.The group method of data handling with differential evolution(GMDH-DE)algorithm was used to predict permeability due to its capability to manage complex,nonlinear relationships between variables,reduced computation time,and parameter optimization through evolutionary algorithms.Using 1953 samples from Gunya-1 and Gunya-2 wells for training and 1563 samples from Gunya-3 for testing,the GMDH-DE outperformed the group method of data handling(GMDH)and random forest(RF)in predicting permeability with higher accuracy and lower computation time.The GMDH-DE achieved an R^(2)of 0.9985,RMSE of 3.157,MAE of 2.366,and ME of 0.001 during training,and for testing,the ME,MAE,RMSE,and R^(2)were 1.3508,12.503,21.3898,and 0.9534,respectively.Additionally,the GMDH-DE demonstrated a 41%reduction in processing time compared to GMDH and RF.The model was also used to predict the permeability of the Mita Gamma well in the Mandawa basin,Tanzania,which lacks core data.Shapley additive explanations(SHAP)analysis identified thermal neutron porosity(TNPH),effective porosity(PHIE),and spectral gamma-ray(SGR)as the most critical parameters in permeability prediction.Therefore,the GMDH-DE model offers a novel,efficient,and accurate approach for fast permeability prediction,enhancing hydrocarbon exploration and production.
基金supported by the National Natural Science Foundation of China Project(No.62302540)please visit their website at https://www.nsfc.gov.cn/(accessed on 18 June 2024).
文摘The methods of network attacks have become increasingly sophisticated,rendering traditional cybersecurity defense mechanisms insufficient to address novel and complex threats effectively.In recent years,artificial intelligence has achieved significant progress in the field of network security.However,many challenges and issues remain,particularly regarding the interpretability of deep learning and ensemble learning algorithms.To address the challenge of enhancing the interpretability of network attack prediction models,this paper proposes a method that combines Light Gradient Boosting Machine(LGBM)and SHapley Additive exPlanations(SHAP).LGBM is employed to model anomalous fluctuations in various network indicators,enabling the rapid and accurate identification and prediction of potential network attack types,thereby facilitating the implementation of timely defense measures,the model achieved an accuracy of 0.977,precision of 0.985,recall of 0.975,and an F1 score of 0.979,demonstrating better performance compared to other models in the domain of network attack prediction.SHAP is utilized to analyze the black-box decision-making process of the model,providing interpretability by quantifying the contribution of each feature to the prediction results and elucidating the relationships between features.The experimental results demonstrate that the network attack predictionmodel based on LGBM exhibits superior accuracy and outstanding predictive capabilities.Moreover,the SHAP-based interpretability analysis significantly improves the model’s transparency and interpretability.
基金support provided by The Science and Technology Development Fund,Macao SAR,China(File Nos.0057/2020/AGJ and SKL-IOTSC-2021-2023)Science and Technology Program of Guangdong Province,China(Grant No.2021A0505080009).
文摘Accurate prediction of shield tunneling-induced settlement is a complex problem that requires consideration of many influential parameters.Recent studies reveal that machine learning(ML)algorithms can predict the settlement caused by tunneling.However,well-performing ML models are usually less interpretable.Irrelevant input features decrease the performance and interpretability of an ML model.Nonetheless,feature selection,a critical step in the ML pipeline,is usually ignored in most studies that focused on predicting tunneling-induced settlement.This study applies four techniques,i.e.Pearson correlation method,sequential forward selection(SFS),sequential backward selection(SBS)and Boruta algorithm,to investigate the effect of feature selection on the model’s performance when predicting the tunneling-induced maximum surface settlement(S_(max)).The data set used in this study was compiled from two metro tunnel projects excavated in Hangzhou,China using earth pressure balance(EPB)shields and consists of 14 input features and a single output(i.e.S_(max)).The ML model that is trained on features selected from the Boruta algorithm demonstrates the best performance in both the training and testing phases.The relevant features chosen from the Boruta algorithm further indicate that tunneling-induced settlement is affected by parameters related to tunnel geometry,geological conditions and shield operation.The recently proposed Shapley additive explanations(SHAP)method explores how the input features contribute to the output of a complex ML model.It is observed that the larger settlements are induced during shield tunneling in silty clay.Moreover,the SHAP analysis reveals that the low magnitudes of face pressure at the top of the shield increase the model’s output。
基金funded by the University Teachers Innovation Fund Project of Gansu Province(2025A-001)the Northwest Normal University Young Teachers'Scientific Research Ability Improvement Plan(NWNULKQN2024-20).
文摘Accurately revealing the spatial heterogeneity in the trade-offs and synergies of land use functions(LUFs)and their driving factors is imperative for advancing sustainable land utilization and optimizing land use planning.This is especially critical for ecologically vulnerable inland river basins in arid regions.However,existing methods struggle to effectively capture complex nonlinear interactions among environmental factors and their multifaceted relationships with trade-offs and synergies of LUFs,especially for the inland river basins in arid regions.Consequently,this study focused on the middle reaches of the Heihe River Basin(MHRB),an arid inland river basin in northwestern China.Using land use,socioeconomic,meteorological,and hydrological data from 2000 to 2020,we analyzed the spatiotemporal patterns of LUFs and their trade-off and synergy relationships from the perspective of production,living,ecological functions.Additionally,we employed an integrated Extreme Gradient Boosting(XGBoost)-SHapley Additive exPlanations(SHAP)framework to investigate the environmental factors influencing the spatial heterogeneity in the trade-offs and synergies of LUFs.Our findings reveal that from 2000 to 2020,the production,living,and ecological functions of land use within the MHRB exhibited an increasing trend,demonstrating a distinct spatial pattern of''high in the southwest and low in the northeast''.Significant spatial heterogeneity defined the trade-off and synergistic relationships,with trade-offs dominating human activity-intensive oasis areas,while synergies prevailed in other areas.During the study period,synergistic relationships between production and living functions and between production and ecological functions were relatively robust,whereas synergies in living-ecological functions remained weaker.Natural factors(digital elevation model(DEM),annual mean temperature,Normalized Difference Vegetation Index(NDVI),and annual precipitation)emerged as the primary factors driving the trade-offs and synergies of LUFs,followed by socioeconomic factors(population density,Gross Domestic Product(GDP),and land use intensity),while distance factors(distance to water bodies,distance to residential areas,and distance to roads)exerted minimal influence.Notably,the interactions among NDVI,annual mean temperature,DEM,and land use intensity exerted the most substantial impacts on the relationships among LUFs.This study provides novel perspectives and methodologies for unraveling the mechanisms underlying the spatial heterogeneity in the trade-offs and synergies of LUFs,offering scientific insights to inform regional land use planning and sustainable natural resource management in inland river basins in arid regions.
基金The National Natural Science Foundation of China (No. 52272367)the Natural Science Foundation of Jiangsu Province (No. BK20231324)。
文摘Ensuring independent mobility for older adults has become a public health and social concern in China owing to its rapidly aging population.To explore independent mobility trends among older adults and the impact of sociodemo-graphic characteristics in recent years,this study used data from the Chinese Longitudinal Healthy Longevity Survey from 2012 to 2018,combined with binomial logit regression and CatBoost-Shapley additive explanation(SHAP)method to analyze the relationship between independent mobility and sociodemographic characteristics under bus and walking-oriented environments.Study findings indicated that age and gender significantly affected the independent mobility of older adults.Policymaking should prioritize the needs of older adults,focusing on age and gender differ-ences.Additionally,living expense adequacy significantly influenced independent mobility.Policies should substan-tially support economically disadvantaged older adults,en-suring their basic needs are met through subsidies and other measures.Moreover,the study found a notable impact of widowhood on independent mobility,suggesting enhanced social care and mental health support for widowed older adults,especially those who are long-lived.The outcomes of this study provided evidence for policymakers,which are beneficial for developing elderly-friendly travel policies to ensure and enhance the quality of life and independent mo-bility of older adults.
基金supported by the National Key R&D Program of China(2023YFE0207900)。
文摘In the arid regions of Northwest China,vegetation cover plays a crucial role in maintaining unique terrestrial ecosystems.Vegetation growth is highly sensitive to variations in topographical factors,and the influence of topography on vegetation cover has attracted increasing attention.This study analyzed vegetation dynamics and their relationship with topography in the Tianshan Mountains of China using Landsat Normalized Difference Vegetation Index(NDVI)data during 2000–2022 and Shuttle Radar Topography Mission(SRTM)-derived topographical factors(elevation,slope,and aspect).Theil-Sen slope estimation and Mann-Kendall trend tests were applied to quantify temporal changes in vegetation,while a terrain area correction coefficient(K)was used to assess spatial associations of vegetation with topography.Random Forest(RF)regression and SHapley Additive exPlanations(SHAP)analysis evaluated the relative importance of topographical factors in shaping vegetation cover(multi-year mean NDVI)distribution.Key findings included that over the 23-a period,59.46%of the vegetated area exhibited significant improvement(P<0.05),with the southern Tianshan Mountains showing the most pronounced increase(70.59%),whereas vegetation degradation(3.10%)was primarily concentrated in river valleys with intensive human activities.RF-SHAP analysis revealed that elevation is the primary driver of vegetation cover patterns,explaining 52.00%of the NDVI variation.The peak NDVI(0.42)occurred at elevations between 2800 and 3200 m.Slope and aspect also significantly influenced vegetation distribution,and higher NDVI values and greater improvement trends were observed on shady(north-facing)slopes compared to sunny(south-facing)slopes.K-index analysis indicated pronounced vegetation change—both degradation and improvement—in areas with elevations between 1100 and 2800 m and slopes exceeding 5°,particularly on sunny slopes.Low-elevation desert areas in the southern Tianshan Mountains were highly susceptible to degradation.This study underscores the critical role of topography in regulating vegetation cover and its spatiotemporal dynamics,providing a scientific basis for sustainable management of arid mountain ecosystems.
基金supported by the Natural Science Foundation of Xinjiang Uygur Autonomous Region(2022D01B109)the Tianchi Doctoral Program of Xinjiang Uygur Autonomous Region(BS2021007).
文摘Within the context of global climate change and rapid urbanization,increasing urban resilience(UR)is especially important in the arid region of Northwest China(ANC),where fragile ecosystems and an uneven water distribution create significant sustainability challenges.In this study,a coupled UR-water ecosystem services(WESs)framework was developed on the basis of 1-km resolution remote sensing data for the 2000–2020 period obtained from the Landsat series,Defense Meteorological Satellite Program(DMSP)/Operational Linescan System(OLS),and Global Precipitation Measurement(GPM),among other sources.Within the framework,the Integrated Valuation of Ecosystem Services and Tradeoffs(InVEST)model was incorporated to provide a WES indicator system.Moreover,entropy weighting was employed to quantify both UR and WES indicators;the coupling coordination degree(CCD)model was used to measure the coupled relationship between UR and WESs;an extreme gradient boosting(XGBoost)-SHapley Additive exPlanations(SHAP)interpretation approach was adopted to identify key drivers and underlying mechanisms;and Geographically Weighted Regression(GWR)was applied to capture spatial distribution characteristics of major driving factors.The results indicated that UR steadily increased from 4.60×10^(-3) to 10.24×10^(-3),whereas WESs followed an inverted V-shaped trend,with a peak value observed in 2010(11.84×10^(-3)).The CCD remained consistently low(mean:0.0166–0.0246)and exhibited considerable spatial heterogeneity.Notably,the degree of coordination was greater in the oasis and mountain core areas but significantly lower at desert areas.XGBoost-SHAP model analysis revealed six key drivers influencing various states of the CCD between UR and WESs systems.The contributions of these factors could be ranked as follows:water yield(WY;24.30%)>farmland area per capita(FP;21.10%)>gross domestic product(GDP)per capita(GDPC;19.00%)>soil retention(SR;14.90%)>population density(PD;5.42%)>water purification(WP;4.40%).In contrast,in UR system,WP(53.66%)and SR(31.62%)served as the dominant drivers.Moreover,the dominant drivers shifted from a combination of natural and socioeconomic factors in StateⅠ(sustainable high resilience)to primarily socioeconomic factors in StateⅢ(unsustainable low resilience).SR and WP exerted positive moderating effects,whereas socioeconomic factors such as GDPC and PD exerted inhibitory effects on the coordination relationship.This research highlights that UR in the ANC region is limited mainly by water scarcity,weak feedback loops,and spatial variability,emphasizing the need for tailored intervention strategies.
基金supported by the Natural Science Foundation of Xinjiang Uygur Autonomous Region(2023E01006,2024TSYCCX0004).
文摘Arid mountain ecosystems are highly sensitive to hydrothermal stress and land use intensification,yet where net primary productivity(NPP)degradation is likely to persist and what drives it remain unclear in the Tianshan Mountains of Northwest China.We integrated multi-source remote sensing with the Carnegie–Ames–Stanford Approach(CASA)model to estimate NPP during 2000–2020,assessed trend persistence using the Hurst exponent,and identified key drivers and nonlinear thresholds with Extreme Gradient Boosting(XGBoost)and SHapley Additive exPlanations(SHAP).Total NPP averaged 55.74 Tg C/a and ranged from 48.07 to 65.91 Tg C/a from 2000 to 2020,while regional mean NPP rose from 138.97 to 160.69 g C/(m^(2)·a).Land use transfer analysis showed that grassland expanded mainly at the expense of unutilized land and that cropland increased overall.Although NPP increased across 64.11%of the region during 2000–2020,persistence analysis suggested that 53.93%of the Tianshan Mountains was prone to continued NPP decline,including 36.41%with significant projected decline and 17.52%with weak projected decline;these areas formed degradation hotspots concentrated in the central and northern Tianshan Mountains.In contrast,potential improvement was limited(strong persistent improvement:4.97%;strong anti-persistent improvement:0.36%).Driver attribution indicated that land use dominated NPP variability(mean absolute SHAP value=29.54%),followed by precipitation(16.03%)and temperature(11.05%).SHAP dependence analyses showed that precipitation effects stabilized at 300.00–400.00 mm,and temperature exhibited an inverted U-shaped response with a peak near 0.00°C.These findings indicated that persistent degradation risk arose from hydrothermal constraints interacting with land use conversion,highlighting the need for threshold-informed,spatially targeted management to sustain carbon sequestration in arid mountain ecosystems.
基金supported by the Key Research and Development Project of Xinjiang Uygur Autonomous Region,China(2022B02049)the Major Science and Technology Special Project of Xinjiang Uygur Autonomous Region,China(2024A03007-5).
文摘In the northern Tarim River Basin,the Weigan River Basin is a critical endorheic system characterized by extreme aridity,where drought poses a major natural hazard to agricultural production and ecological stability.This study assessed the future evolution of drought under climate change by employing the standardized moisture anomaly index(SZI)on the basis of multi-model the Coupled Model Intercomparison Project Phase 6(CMIP6)simulations under historical conditions(1970–2014)and future scenarios(shared socioeconomic pathway(SSP)1-2.6,SSP2-4.5,SSP3-7.0 and SSP5-8.5 for 2015–2100).The results show that precipitation–evapotranspiration anomalies are projected to first decline but then increase over time,with increased fluctuations and uncertainty under high-emission scenarios(SSP5-8.5).These trends indicate intensifying drought risks and reveal a strong influence of emission pathways on regional water cycling.Temporal analysis of SZI indicates a transition from wetting to drying under lowand medium-emission pathways(SSP1-2.6 and SSP2-4.5),whereas high-emission scenarios are characterized by persistent drying and increased variability.The significant lower-tail dependence(0.271)observed under SSP2-4.5 and SSP5-8.5 suggests that extreme droughts may be subject to nonlinear co-amplification across scenarios.The frequency of moderate and more severe drought events is expected to increase substantially,especially under SSP5-8.5,where drought occurrence is predicted to extend into spring and autumn and become more evenly distributed throughout the year.Spatially,drought duration shows significant positive autocorrelation across all scenarios,with hot spots consistently concentrated in the southern and southeastern regions of the basin.Random forest analysis,interpreted as association-based pattern attribution,indicates that meteorological variables(precipitation and potential evapotranspiration(PET))make the greatest contributions to the hot spot pattern,followed by topography and soil moisture.Among land use categories,farmland generally shows higher drought sensitivity than other land use types,as reflected by its relative contribution patterns across scenarios.The spatial pattern of drought is statistically structured by climatic forcing,surface conditions,and soil moisture status,reflecting their coupled associations with hot spot occurrence.In addition,a drought spatial uncertainty index was constructed from multi-scenario hot spot maps,revealing spatially heterogeneous structural variability throughout the basin.Correlation analysis further highlights strong internal couplings among environmental variables(e.g.,elevation-linked hydroclimatic gradients and grassland–bare soil contrasts).These findings offer a scientific basis for developing region-specific drought monitoring and adaptation strategies under future climate change conditions.
基金supported by the Young Mentor Fund project of Gansu Agricultural University[grant number 0522014]the Gansu Provincial Science and Technology Plan[grant number 23CXNA0017].
文摘Color has emerged as a pivotal factor influencing consumer purchasing decisions in the dried herbal medicine market.To address the issue of significant discoloration of Rhubarb(Rheum rhabarbarum L.)during the drying process,this study investigates the effects of microwave fixation(MF)and hot-air fixation(HAF)pretreatment methods on the drying characteristics and quality of Rhubarb by ultrasonic synergistic vacuum far-infrared drying(U-VFID),with a primary focus on its color attributes.The results indicate that fixation pretreatment significantly enhances both drying efficiency and product quality,particularly in terms of color retention.Compared to unpretreated Rhubarb,the best comprehensive drying effect was achieved with MF treatment at 60℃for 7 min,which resulted in a 441.18%increase in rhein content,a 58.57%reduction in drying time,and a 48.38%decrease in theΔE value.Furthermore,correlation analysis,and the eXtreme Gradient Boosting(XGBoost)algorithm combined with SHapley Additive exPlanations(SHAP)model,revealed that the color of Rhubarb subjected to various fixation pretreatments in conjunction with U-VFID is primarily influenced by sennoside A content,total phenolic content(TPC),and drying time.This study offers a scientific foundation and theoretical insights for optimizing the quality of dried medicinal plant products and introduces innovative approaches for post-harvest industrial pretreatment of rhizomatous medicinal plants.
文摘本文聚焦我国青海湖流域的水文过程,基于多年气象和水文动态数据,发展了一种融合概念性水文模型FLEX (FluxExchange)和门控循环单元(Gated Recurrent Unit, GRU)的混合模型对流域内最大支流布哈河的逐日径流进行了模拟和预测。在构建混合模型中,采用了三种策略提升模拟精度:引入差分进化自适应算法DREAM(zs)反演水文参数优化FLEX模型;采用变分模态分解(VMD)提取径流数据的信息和特征;利用麻雀搜索算法(SSA)优化深度学习GRU的参数。研究将FLEX模型的模拟结果连同气象数据一起作为神经网络的输入,从而构建了FLEX-VMD-SSA-GRU混合模型。同时,探讨了不同的气象输入条件对模拟结果的影响和贡献:基于7个主要气象要素,由少及多设置了14组输入情景模拟。最后通过SHAP对深度学习方法的结果进行分析,揭示了气象变量对径流长期趋势的贡献和重要度。
基金supported by the National Key Research and Development Program of China(Grant No.2023YFC3007203).
文摘This study addresses gaps in aftershock prediction research by proposing an interpretable hybrid machine learning model that leverages multi-source data.The model overcomes challenges related to the selection of influencing factors,model types,prediction result visualization,and decision mechanism interpretability.It integrates mainshock factors,geological features,site characteristics,and terrain conditions using geospatial information system(GIS)technology.By employing the stacking algorithm to optimize and combine XGBoost and LightGBM models,the proposed model significantly improves the prediction performance.Visualization through aftershock hazard mapping offers a robust tool for aftershock warning.The Shapley additive explanations(SHAP)model is used to explain the decision-making process from both global and local perspectives.Results show that,compared to the optimized XGBoost-CMA_ES and LightGBM-CMA_ES hybrid models,the stacking model achieves area under the curve(AUC)increases of 7.71%and 5.72% on the test set,respectively,with a maximum prediction accuracy of 0.9344.The hazard zoning map identifies high-risk areas mainly around fault lines and near the epicenter.As hazard levels rise,the proportion and density of aftershocks in these areas increase.The SHAP model results highlight the distance to fault as the most critical factor.The study integrates local explanations with on-site investigations,effectively visualizing the contributions of different factors to aftershocks.This research provides new tools and methods for enhancing aftershock warning and response.
文摘Shield attitudes,essentially governed by intricate mechanisms,impact the segment assembly quality and tunnel axis deviation.In data-driven prediction,however,existing methods using the original driving parameters fail to present convincing performance due to insufficient consideration of complicated interactions among the parameters.Therefore,a multi-dimensional feature synthesizing and screening method is proposed to explore the optimal features that can better reflect the physical mechanism in predicting shield tunneling attitudes.Features embedded with physical knowledge were synthesized from seven dimensions,which were validated by the clustering quality of Shapley Additive Explanations(SHAP)values.Subsequently,a novel index,Expected Impact Index(EII),has been proposed for screening the optimal features reliably.Finally,a Bayesian-optimized deep learning model was established to validate the proposed method in a case study.Results show that the proposed method effectively identifies the optimal parameters for shield attitude prediction,with an average Mean Squared Error(MSE)deduction of 27.3%.The proposed method realized effective assimilation of shield driving data with physical mechanism,providing a valuable reference for shield deviation control.
基金Supported by the National Natural Science Foundation of China(No.42276019)the Guangdong Provincial Observation and Research Station for Tropical Ocean Environment in Western Coastal Waters(No.GSTOEW)。
文摘The uncertainty of ocean turbulent mixing parameterization comprises a significant challenge in ocean and climate models. A depth-dependent deep learning ocean turbulent mixing parameterization scheme was proposed with the hydrological and microstructure observations conducted in summer 2012 in the shelf sea east of Hainan Island, in South China Sea(SCS). The deep neural network model is used and incorporates the Richardson number Ri, the normalized depth D, the horizontal velocity speed U, the shear S^(2), the stratification N^(2), and the density ρ as input parameters. Comparing to the scheme without parameter D and region division, the depth-dependent scheme improves the prediction of the turbulent kinetic energy dissipation rate ε. The correlation coefficient(r) between predicted and observed lgε increases from 0.49 to 0.62, and the root mean square error decreases from 0.56 to 0.48. Comparing to the traditional physics-driven parameterization schemes, such as the G89 and MG03, the data-driven approach achieves higher accuracy and generalization. The SHapley Additive Explanations(SHAP) framework analysis reveals the importance descending order of the input parameters as: ρ, D, U, N^(2), S^(2), and Ri in the whole depth, while D is most important in the upper and bottom boundary layers(D≤0.3&D≥0.65) and least important in middle layer(0.3<D<0.65). The research shows applicability of constructing deep learning-based ocean turbulent mixing parameterization schemes using limited observational data and well-established physical processes.
基金support from the National Key Research and Development Program of China(Grant Nos.2023YFC3008300 and 2023YFC3008305)the National Natural Science Foundation of China(Grant Nos.42172296).
文摘Accurate and interpretable prediction of shield tunnel-induced settlement poses a significant challenge due to the complex interplay of various influencing factors.This paper proposes a novel interpretable hybrid model that combines complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN),slime mold algorithm(SMA),and least squares support vector machine(LSSVM)to enhance prediction accuracy and model transparency.The CEEMDAN method,optimized by SMA,decomposes settlement data into intrinsic mode functions(IMFs)and residuals,thereby reducing data noise.The LSSVM,also optimized by SMA,is then applied to predict each IMF and residual.The final settlement prediction is derived from the aggregation of these results.The model was rigorously validated using the Changsha(China)and Singapore Metro projects,demonstrating superior performance to traditional machine learning models.The evaluation metrics,including root mean square error(RMSE),mean absolute error(MAE),mean absolute percentage error(MAPE),and coefficient of determination(R2),underscore the model's effectiveness.The model achieved the lowest error rates and highest accuracy across these metrics.Notably,Shapley additive explanations(SHAP)provided insights into the model's decision-making process,identifying shield stoppage and moisture content as the most influential factors in settlement prediction.This study contributes to the advancement of the methodological framework for predicting tunnel settlement.It addresses the discrepancy between prediction accuracy and interpretability,providing a robust tool for practical engineering applications.
