A Support Vector Machine and Particle Swarm Optimization Based Model for Cemented Tailings Backfill Materials...

Area of Science:

• Geotechnical Engineering
• Computational Intelligence
• Materials Science

Background:

• Accurate prediction of unconfined compressive strength (UCS) is crucial for cemented paste backfill (CTB) performance.
• Optimization algorithms and machine learning offer potential for enhancing predictive accuracy.

Purpose of the Study:

• To investigate the feasibility of a hybrid model combining Particle Swarm Optimization (PSO) and Support Vector Machine (SVM) for predicting CTB UCS.
• To establish an accurate and efficient predictive model for CTB UCS using experimental data.

Main Methods:

• Dataset construction based on experimental UCS values.
• Application of categorized random segmentation for training set establishment.
• Hyperparameter tuning of SVM using PSO, identifying optimal parameters (C=71.923, ε=0.0625, γ=0.195).

Main Results:

• The PSO-SVM model demonstrated high accuracy and efficiency in UCS prediction.
• Achieved a coefficient of determination (R²) of 0.97 and a Mean Squared Error (MSE) of 0.0044.
• Categorized random segmentation proved suitable for training set creation.

Conclusions:

• The developed PSO-SVM model is feasible and effective for predicting CTB UCS.
• The model's accuracy and robustness are expected to improve with larger datasets.
• This approach offers a reliable tool for CTB engineering applications.