Explainable Artificial Intelligence Model for Evaluating Shear Strength Parameters of Municipal Solid Waste Across Diverse Compositional Profiles

Transparency in AI improves engineering decisions: The study demonstrates that explainable AI (using SHAP analysis) can reveal why a model makes predictions about waste strength, showing that fibrous materials and particle size are the primary factors—information that helps engineers make evidence-based decisions rather than relying on "black box" predictions.

AI accuracy enables safer waste management infrastructure: The model achieved 7.42% error rates for friction angle predictions, substantially outperforming traditional methods. This level of accuracy is critical for designing safer landfills and waste facilities, where miscalculations could lead to structural failures.

Understanding non-linear relationships prevents oversimplification: The research revealed that food waste and plastics have "significant but non-linear effects" on shear strength, meaning simple rules-of-thumb would fail. The XAI framework captures these complex interactions that traditional engineering approaches might miss.

Explainability builds trust for real-world adoption: By quantifying how specific waste components influence structural properties, the model provides the transparency needed for engineers and regulators to confidently adopt AI recommendations in practice, rather than viewing them as untrustworthy algorithmic outputs.

Responsible AI requires understanding model limitations: The study's focus on explaining which factors matter most helps practitioners recognize when predictions may be unreliable—for example, when waste compositions fall outside the training data's diversity—supporting more cautious and responsible deployment.