Numerical model of debris flow susceptibility using slope stability failure machine learning prediction with...

Area of Science:

• Geotechnical Engineering
• Computational Science
• Natural Hazard Assessment

Background:

• Slope stability analysis is crucial for debris flow hazard assessment.
• Traditional field studies for debris flow monitoring are costly and time-consuming.
• Developing intelligent models can reduce costs and improve efficiency in slope behavior analysis.

Purpose of the Study:

• To develop intelligent numerical models for predicting debris flow susceptibility.
• To enhance machine learning model performance using novel metaheuristic training methods.
• To provide a cost-effective and time-efficient approach for slope design and monitoring.

Main Methods:

• Development of intelligent numerical models for Factor of Safety (FOS) prediction.
• Application of novel metaheuristic methods for training machine learning models.
• Utilizing adaptive neuro-fuzzy inference system (ANFIS) combined with particle swarm optimization (PSO).

Main Results:

• The ANFIS-PSO model demonstrated superior performance in predicting debris flow FOS.
• Achieved over 85% accuracy in FOS prediction, outperforming other tested methods.
• Validated model accuracy using multiple performance evaluation indices.

Conclusions:

• Intelligent models offer a robust and efficient alternative for debris flow susceptibility prediction.
• Metaheuristic training significantly enhances machine learning model performance.
• The developed ANFIS-PSO model provides a reliable tool for cost-effective slope stability assessment.