Predicting waste shear strength through explainable triple-stream attention deep learning

Abstract

Sustainable waste management requires accurate characterization of municipal solid waste (MSW), which remains a critical challenge in geotechnical engineering due to the heterogeneous nature of waste materials and their temporal evolution through degradation processes. This paper presents a novel explainable artificial intelligence (XAI) framework for evaluating cohesion and friction angle across diverse MSW compositional profiles. The proposed Triple-Stream Attention Network (TSAN) combines feature attention, batch attention, and a residual multi-layer perceptron stream, with SHAP (SHapley Additive exPlanations) applied as a post-hoc interpretability layer to provide transparent insights into how specific waste components influence strength characteristics. Training data encompassed large-scale direct shear tests across various waste compositions and degradation states. The model demonstrated superior predictive accuracy compared to traditional gradient boosting methods, achieving mean absolute percentage errors of 2.66% and 9.78% for friction angle and cohesion predictions, respectively. Through SHAP analysis, the study revealed that fibrous materials and particle size distribution were primary drivers of shear strength variation, with food waste and plastics showing significant but non-linear effects. The model’s explainability component successfully quantified these relationships, enabling evidence-based recommendations for waste management practices. Notably, the dataset derives from a single large-scale dumpsite and captures degradation cross-sectionally (space-for-time) rather than through longitudinal monitoring, which constitute acknowledged limitations. This research bridges the gap between advanced machine learning and geotechnical engineering practice, offering a reliable tool for rapid assessment of MSW mechanical properties while maintaining interpretability for engineering decision-making.