Stiffness and Strength of Stabilized Organic Soils—Part II/II: Parametric Analysis and Modeling with Machine Learning

Predicting the range of achievable strength and stiffness from stabilized soil mixtures is critical for engineering design construction, especially organic soils, which are often considered “unsuitable” due to their high compressibility lack knowledge about mechanical behavior after stabilization. This study investigates soils using machine learning (ML) methods. ML algorithms were developed trained a database comprehensive experimental (see Part I), including more than one thousand unconfined compression tests on clay samples by wet mixing (WSM) technique. Three different methods adopted compared, two artificial neural networks (ANN) linear regression method. ANN models proved reliable in prediction significantly outperforming models. Binder type, ratio, water content, sample size, aging, temperature, relative humidity, carbonation control variables (input parameters) incorporated into The impacts these factors evaluated through rigorous ANN-based parametric analyses. Additionally, nonlinear relations with parameters developed, optimum ranges identified Overall, robust approach presented this paper can improve mixture stabilization minimize cost implementing WSM projects.