Machine Learning in Geohazard Risk Prediction and Assessment : From Microscale Analysis to Regional Mapping

• Record Nr.: UNINA9911044026103321
• Autore: Pradhan Biswajeet
• Pubbl/distr/stampa: Chantilly : , : Elsevier, , 2025; ©2025
• ISBN: 0-443-23664-X
• Edizione: [1st ed.]
• Descrizione fisica: 1 online resource (377 pages)
• Altri autori (Persone): ShengDaichao; HeXuzhen
• Disciplina: 550.285631
• Lingua di pubblicazione: Inglese
• Formato: Materiale a stampa
• Livello bibliografico: Monografia
• Nota di contenuto: Front Cover -- Machine Learning in Geohazard Risk Prediction and Assessment -- Copyright Page -- Contents -- List of contributors -- About the editors -- Preface -- 1 Machine learning methods and connections between different parts -- 1 Machine learning methods and their connections -- 1.1 Introduction -- 1.2 Overview of machine learning methods -- 1.2.1 Supervised learning -- 1.2.1.1 Linear regression -- 1.2.1.2 Logistic regression -- 1.2.1.3 K-nearest neighbors -- 1.2.1.4 Support vector machine -- 1.2.1.5 Decision tree -- 1.2.1.6 Random forest -- 1.2.1.7 Artificial neural network -- 1.2.1.8 Convolutional neural network -- 1.2.1.9 Recurrent neural network -- 1.2.2 Semi-supervised learning -- 1.2.2.1 Label propagation -- 1.2.2.2 Self-training -- 1.2.3 Unsupervised learning -- 1.2.3.1 K-means clustering -- 1.2.3.2 Hierarchical clustering -- 1.2.3.3 Principal component analysis -- 1.2.3.4 Autoencoder -- 1.2.4 Reinforcement learning -- 1.2.4.1 Q-learning -- 1.2.4.2 Deep Q networks -- 1.2.4.3 Policy gradients -- 1.2.5 Others -- 1.3 Performance evaluation metrics -- 1.3.1 Pearson correlation coefficient (R) -- 1.3.2 Coefficient of determination (R2) -- 1.3.3 Mean square error -- 1.3.4 Root mean square error -- 1.3.5 Mean absolute error -- 1.3.6 Mean absolute percentage error -- 1.3.7 Variance account for -- 1.4 Machine learning performance optimization techniques -- 1.4.1 Data preprocessing -- 1.4.2 Generalization enhancement -- 1.5 Optimization algorithms -- ; 1.5.1 Particle swarm optimization -- 1.5.2 Ant colony optimization -- 1.5.3 Bat algorithm -- 1.5.4 Grey wolf optimizer -- 1.5.5 Artificial fish swarm algorithm -- 1.5.6 Whale optimization algorithm -- 1.5.7 Genetic algorithm -- 1.5.8 Simulated annealing -- 1.6 The connections and challenges among different machine learning methods -- 1.6.1 Connections -- 1.6.2 Challenges.; 1.6.3 Future development -- 1.7 Conclusion -- References -- 2 Machine learning in constitutive modelling of geo-materials -- 2 Analysis and quantitative identification of tensile and shear fractures in rocks: a review and prospect -- 2.1 Introduction -- 2.2 Method of characterizing rock fracture mechanism -- 2.2.1 Acoustic emission -- 2.2.2 Three-dimensional optical morphology scanning -- 2.2.3 Digital image correlation technique -- 2.2.4 Scanning electron microscopy -- 2.3 Quantitative identification of mesoscopic fracture mechanism -- 2.3.1 Experiment tests -- 2.3.2 SEM images acquisition -- 2.3.3 Modelling -- 2.4 Discussion -- 2.4.1 Model comparison -- 2.4.2 Model interpretation -- 2.5 Model feasibility validation -- 2.6 Conclusions -- References -- 3 Tree-based machine-learning models to estimate undrained shear strength of soft clay soils -- 3.1 Introduction -- 3.2 Undrained shear strength -- 3.3 Variable parameters -- 3.4 Gradient boosting tree and extreme gradient boosting tree -- 3.5 Data analysis -- 3.6 Results and discussion -- 3.7 Conclusion -- References -- 4 Predicting liquefaction resistance of silty sands using machine learning techniques -- 4.1 Introduction -- 4.2 Liquefaction fundamentals -- 4.2.1 Seismic hazards and liquefaction-induced damage -- 4.2.2 Understanding liquefaction -- 4.3 Mechanics of liquefaction -- 4.4 Factors Influencing ground motions -- 4.4.1 Ground failures induced by liquefaction -- 4.4.1.1 Flow failures -- 4.4.1.2 Lateral spreading -- 4.4.1.3 Ground oscillation -- 4.4.1.4 Loss of bearing capacity -- 4.5 Factors affecting liquefaction susceptibility -- 4.5.1 Grain-size distribution and soil types -- 4.5.2 Relative density -- 4.5.3 Earthquake loading characteristics -- 4.5.4 Vertical effective stress and over consolidation -- 4.5.5 Soil age and origin -- 4.5.6 Seismic strain history -- 4.5.7 Degree of saturation.; 4.5.8 Thickness of sand layer -- 4.6 Methods for assessing liquefaction potential -- 4.7 Machine learning methods -- 4.7.1 Back propagation neural network -- 4.7.2 Support vector machine -- 4.7.3 Radial basis function neural network -- 4.8 Methodology -- 4.9 Conclusion -- References -- 3 Machine learning in numerical modelling of geotechnical problems -- 5 Introduction to system reliability analysis for geotechnical infrastructures -- 5.1 Introduction -- 5.2 Basic statistical concepts -- 5.2.1 Random variable -- 5.2.2 Mean of a random variable -- 5.2.3 Standard deviation of a random variable -- 5.2.4 Variance of a random variable -- 5.2.5 Independent random variables -- 5.2.6 Covariance -- 5.2.7 Correlation coefficient -- 5.2.8 Probability density function and cumulative distribution function -- 5.2.9 Probability distribution functions -- 5.2.9.1 Normal distribution function -- 5.2.9.2 Lognormal distribution function -- 5.2.9.3 Probability distribution function -- 5.2.9.4 Exponential distribution function -- 5.2.9.5 Beta distribution function -- 5.3 Reliability assessment methods -- 5.3.1 Fundamental reliability concepts -- 5.3.1.1 Central safety factor, reliability index, and probability of failure (Pf) -- 5.3.2 Monte Carlo simulation -- 5.3.3 First order reliability method -- 5.3.3.1 Illustrative example 1 -- 5.4 Series, parallel, and combined systems -- 5.4.1 Reliability of series systems -- 5.4.2 Reliability of parallel systems -- 5.4.3 Cornell bound method -- 5.4.4 Reliability of combined systems -- 5.4.5 Equivalent linear safety margin for parallel systems -- 5.4.5.1 Illustrative example 2 -- 5.5 Sequential ; compounding method -- 5.5.1 Compounding two components coupled by intersection -- 5.5.2 Compounding two components coupled by union -- 5.5.2.1 Illustrative example 3 -- 5.6 Summary -- 5.7 Conclusion -- References.; 6 Deep learning for surrogate modeling for geotechnical risk analysis -- 6.1 Introduction -- 6.2 Stability analysis methods -- 6.2.1 Limit equilibrium method and the limit analysis method -- 6.2.2 Finite element method -- 6.3 Homogeneous slope stability -- response surface method -- 6.3.1 The Limit analysis method for slope stability analysis -- 6.3.2 A case study of slope toe failure -- 6.3.3 Stochastic analysis for slope stability by response surface method -- 6.4 Spatially variable slopes -- machine learning surrogate model -- 6.4.1 Slope stability analysis by finite element method -- 6.4.2 Stochastic slope stability analysis -- 6.4.3 Machine learning surrogate models for slope stability analysis -- 6.5 Spatially variable slopes of any geometries or material properties -- deep-learning pretrained surrogate model -- 6.5.1 Slope stability: inputs and outputs -- 6.5.2 FE models and calculation of safety factors -- 6.5.3 Deep-learning models for spatially variable slopes -- 6.5.3.1 Undrained soil with a fixed slope shape -- 6.5.3.2 Mohr-Coulomb soil with a fixed slope shape -- 6.5.3.3 The full problem -- 6.5.4 Verification and application -- 6.6 Conclusion -- 6.7 AI disclosure -- References -- 7 Physics-based deep-learning numerical methods: application to groundwater seepage and consolidation problems -- 7.1 Introduction -- 7.2 Main theories of PBDL algorithms -- 7.2.1 Principle of PINNs -- 7.2.2 Principle of PI-DeepONets -- 7.3 Numerical applications of PBDL algorithms -- 7.3.1 Forward nonlinear consolidation problem via the PINNs -- 7.3.2 Inverse unsaturated seepage problem via the PINNs -- 7.3.3 Uncertain saturated seepage problem via the PI-DeepONets -- 7.4 Conclusions -- Acknowledgments -- References -- 4 Machine learning in data-drivin geotechnical engineering -- 8 Deep learning for time series forecasting in tunneling -- 8.1 Introduction.; 8.2 Data sampling -- 8.2.1 high-frequency data -- 8.2.2 Low-frequency data -- 8.2.3 Data preprocessing -- 8.3 Application in TBM operation -- 8.3.1 Forecasting TBM performance -- 8.3.1.1 High-frequency forecast -- 8.3.1.2 Low-frequency forecast -- 8.3.1.3 High-to-low forecast -- 8.3.2 Forecasting attitude and position -- 8.3.3 Forecasting tunnel collapse -- 8.4 Summary -- 8.5 Perspectives -- 8.6 AI disclosure -- References -- 9 Computational intelligence in rock engineering management systems -- 9.1 Introduction -- 9.2 Computational intelligence methods -- 9.2.1 Artificial neural network -- 9.2.2 Fuzzy logic -- 9.2.3 Metaheuristic optimization -- 9.2.4 Hybrid method -- 9.2.5 Challenges in applying computational intelligence -- 9.3 Application of computational intelligence frameworks in rock engineering management systems -- 9.3.1 Recognition and assessment of initial conditions -- 9.3.2 Monitoring and surveillance -- 9.3.3 Performance prediction and improvement -- 9.4 Conclusion -- References -- 10 Genetic programming for the prediction of tunnel face support pressure -- 10.1 Introduction -- 10.2 Methods -- 10.2.1 Tunnel face stability -- 10.2.2 Data acquisition -- 10.2.3 Genetic programming -- 10.2.4 Model set up -- 10.3 Results -- 10.3.1 Sensitivity analysis -- 10.4 Discussion -- 10.5 Conclusions -- References -- 11 Machine learning in geo-risk susceptibility mapping -- 11.1 Introduction -- 11.2 Description of study area -- 11.3 Established database -- 11.4 Methodology -- 11.4.1 Artificial neural network -- 11.4.2 Hybrid model development -- 11.4.2.1 Backtracking search algorithm -- 11.4.2.1.1 Initialization -- 11.4.2.1.2 Selection-I -- 11.4.2.1.3 Crossover -- 11.4.2.1.4 Selection-II -- 11.4.2.2 ; Earthworm optimization algorithm -- 11.4.3 Vortex search algorithm -- 11.4.3.1 Whale optimization algorithm (WOA) -- 11.5 Results and discussion.; 11.5.1 Error analysis.
• Sommario/riassunto: Machine Learning in Geohazard Risk Prediction and Assessment: From Microscale Analysis to Regional Mapping presents an overview of the most recent developments in machine learning techniques that have reshaped our understanding of geo-materials and management protocols of geo-risk.